gecko-dev/widget/android/EventDispatcher.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

1097 строки
34 KiB
C++
Исходник Обычный вид История

Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
/* -*- Mode: c++; c-basic-offset: 4; tab-width: 20; indent-tabs-mode: nil; -*-
* vim: set sw=4 ts=4 expandtab:
* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#include "EventDispatcher.h"
#include "JavaBuiltins.h"
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
#include "nsAppShell.h"
#include "nsIXPConnect.h"
#include "nsJSUtils.h"
#include "xpcpublic.h"
#include "mozilla/ScopeExit.h"
#include "mozilla/dom/ScriptSettings.h"
namespace mozilla {
namespace widget {
namespace detail {
bool
CheckJS(JSContext* aCx, bool aResult)
{
if (!aResult) {
JS_ClearPendingException(aCx);
}
return aResult;
}
nsresult
BoxString(JSContext* aCx, JS::HandleValue aData, jni::Object::LocalRef& aOut)
{
if (aData.isNullOrUndefined()) {
aOut = nullptr;
return NS_OK;
}
MOZ_ASSERT(aData.isString());
JS::RootedString str(aCx, aData.toString());
if (JS_StringHasLatin1Chars(str)) {
nsAutoJSString autoStr;
NS_ENSURE_TRUE(CheckJS(aCx, autoStr.init(aCx, str)), NS_ERROR_FAILURE);
// StringParam can automatically convert a nsString to jstring.
aOut = jni::StringParam(autoStr, aOut.Env());
return NS_OK;
}
// Two-byte string
JNIEnv* const env = aOut.Env();
const char16_t* chars;
{
JS::AutoCheckCannotGC nogc;
size_t len = 0;
chars = JS_GetTwoByteStringCharsAndLength(aCx, nogc, str, &len);
if (chars) {
aOut = jni::String::LocalRef::Adopt(env,
env->NewString(reinterpret_cast<const jchar*>(chars), len));
}
}
if (NS_WARN_IF(!CheckJS(aCx, !!chars) || !aOut)) {
env->ExceptionClear();
return NS_ERROR_FAILURE;
}
return NS_OK;
}
nsresult
BoxObject(JSContext* aCx, JS::HandleValue aData, jni::Object::LocalRef& aOut);
template<typename Type,
bool (JS::Value::*IsType)() const,
Type (JS::Value::*ToType)() const,
class ArrayType,
typename ArrayType::LocalRef (*NewArray)(const Type*, size_t)> nsresult
BoxArrayPrimitive(JSContext* aCx, JS::HandleObject aData,
jni::Object::LocalRef& aOut, size_t aLength,
JS::HandleValue aElement)
{
JS::RootedValue element(aCx);
auto data = MakeUnique<Type[]>(aLength);
data[0] = (aElement.get().*ToType)();
for (size_t i = 1; i < aLength; i++) {
NS_ENSURE_TRUE(CheckJS(aCx, JS_GetElement(aCx, aData, i, &element)),
NS_ERROR_FAILURE);
NS_ENSURE_TRUE((element.get().*IsType)(), NS_ERROR_INVALID_ARG);
data[i] = (element.get().*ToType)();
}
aOut = (*NewArray)(data.get(), aLength);
return NS_OK;
}
template<class Type,
nsresult (*Box)(JSContext*, JS::HandleValue, jni::Object::LocalRef&),
typename IsType>
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
nsresult
BoxArrayObject(JSContext* aCx, JS::HandleObject aData,
jni::Object::LocalRef& aOut, size_t aLength,
JS::HandleValue aElement,
IsType&& aIsType)
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
{
auto out = jni::ObjectArray::New<Type>(aLength);
JS::RootedValue element(aCx);
jni::Object::LocalRef jniElement(aOut.Env());
nsresult rv = (*Box)(aCx, aElement, jniElement);
NS_ENSURE_SUCCESS(rv, rv);
out->SetElement(0, jniElement);
for (size_t i = 1; i < aLength; i++) {
NS_ENSURE_TRUE(CheckJS(aCx, JS_GetElement(aCx, aData, i, &element)),
NS_ERROR_FAILURE);
NS_ENSURE_TRUE(element.isNullOrUndefined() || aIsType(element),
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
NS_ERROR_INVALID_ARG);
rv = (*Box)(aCx, element, jniElement);
NS_ENSURE_SUCCESS(rv, rv);
out->SetElement(i, jniElement);
}
aOut = out;
return NS_OK;
}
nsresult
BoxArray(JSContext* aCx, JS::HandleObject aData, jni::Object::LocalRef& aOut)
{
uint32_t length = 0;
NS_ENSURE_TRUE(CheckJS(aCx, JS_GetArrayLength(aCx, aData, &length)),
NS_ERROR_FAILURE);
if (!length) {
// Always represent empty arrays as an empty boolean array.
aOut = java::GeckoBundle::EMPTY_BOOLEAN_ARRAY();
return NS_OK;
}
// We only check the first element's type. If the array has mixed types,
// we'll throw an error during actual conversion.
JS::RootedValue element(aCx);
NS_ENSURE_TRUE(CheckJS(aCx, JS_GetElement(aCx, aData, 0, &element)),
NS_ERROR_FAILURE);
if (element.isBoolean()) {
return BoxArrayPrimitive<bool, &JS::Value::isBoolean, &JS::Value::toBoolean,
jni::BooleanArray, &jni::BooleanArray::New>(
aCx, aData, aOut, length, element);
}
if (element.isInt32()) {
nsresult rv = BoxArrayPrimitive<
int32_t, &JS::Value::isInt32, &JS::Value::toInt32,
jni::IntArray, &jni::IntArray::New>(aCx, aData, aOut,
length, element);
if (rv != NS_ERROR_INVALID_ARG) {
return rv;
}
// Not int32, but we can still try a double array.
}
if (element.isNumber()) {
return BoxArrayPrimitive<
double, &JS::Value::isNumber, &JS::Value::toNumber,
jni::DoubleArray, &jni::DoubleArray::New>(aCx, aData, aOut,
length, element);
}
if (element.isNullOrUndefined() || element.isString()) {
const auto isString = [] (JS::HandleValue val) -> bool {
return val.isString();
};
nsresult rv = BoxArrayObject<jni::String, &BoxString>(
aCx, aData, aOut, length, element, isString);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
if (element.isString() || rv != NS_ERROR_INVALID_ARG) {
return rv;
}
// First element was null/undefined, so it may still be an object array.
}
const auto isObject = [aCx] (JS::HandleValue val) -> bool {
if (!val.isObject()) {
return false;
}
bool array = false;
JS::RootedObject obj(aCx, &val.toObject());
// We don't support array of arrays.
return CheckJS(aCx, JS_IsArrayObject(aCx, obj, &array)) && !array;
};
if (element.isNullOrUndefined() || isObject(element)) {
return BoxArrayObject<java::GeckoBundle, &BoxObject>(
aCx, aData, aOut, length, element, isObject);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
}
NS_WARNING("Unknown type");
return NS_ERROR_INVALID_ARG;
}
nsresult
BoxValue(JSContext* aCx, JS::HandleValue aData, jni::Object::LocalRef& aOut);
nsresult
BoxObject(JSContext* aCx, JS::HandleValue aData, jni::Object::LocalRef& aOut)
{
if (aData.isNullOrUndefined()) {
aOut = nullptr;
return NS_OK;
}
MOZ_ASSERT(aData.isObject());
JS::Rooted<JS::IdVector> ids(aCx, JS::IdVector(aCx));
JS::RootedObject obj(aCx, &aData.toObject());
bool isArray = false;
if (CheckJS(aCx, JS_IsArrayObject(aCx, obj, &isArray)) && isArray) {
return BoxArray(aCx, obj, aOut);
}
NS_ENSURE_TRUE(CheckJS(aCx, JS_Enumerate(aCx, obj, &ids)),
NS_ERROR_FAILURE);
const size_t length = ids.length();
auto keys = jni::ObjectArray::New<jni::String>(length);
auto values = jni::ObjectArray::New<jni::Object>(length);
// Iterate through each property of the JS object.
for (size_t i = 0; i < ids.length(); i++) {
const JS::RootedId id(aCx, ids[i]);
JS::RootedValue idVal(aCx);
JS::RootedValue val(aCx);
jni::Object::LocalRef key(aOut.Env());
jni::Object::LocalRef value(aOut.Env());
NS_ENSURE_TRUE(CheckJS(aCx, JS_IdToValue(aCx, id, &idVal)),
NS_ERROR_FAILURE);
JS::RootedString idStr(aCx, JS::ToString(aCx, idVal));
NS_ENSURE_TRUE(CheckJS(aCx, !!idStr), NS_ERROR_FAILURE);
idVal.setString(idStr);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
NS_ENSURE_SUCCESS(BoxString(aCx, idVal, key), NS_ERROR_FAILURE);
NS_ENSURE_TRUE(CheckJS(aCx, JS_GetPropertyById(aCx, obj, id, &val)),
NS_ERROR_FAILURE);
nsresult rv = BoxValue(aCx, val, value);
if (rv == NS_ERROR_INVALID_ARG && !JS_IsExceptionPending(aCx)) {
nsAutoJSString autoStr;
if (CheckJS(aCx, autoStr.init(aCx, idVal.toString()))) {
JS_ReportErrorUTF8(
aCx, u8"Invalid event data property %s",
NS_ConvertUTF16toUTF8(autoStr).get());
}
}
NS_ENSURE_SUCCESS(rv, rv);
keys->SetElement(i, key);
values->SetElement(i, value);
}
aOut = java::GeckoBundle::New(keys, values);
return NS_OK;
}
nsresult
BoxValue(JSContext* aCx, JS::HandleValue aData, jni::Object::LocalRef& aOut)
{
if (aData.isNullOrUndefined()) {
aOut = nullptr;
} else if (aData.isBoolean()) {
aOut = aData.toBoolean() ? java::sdk::Boolean::TRUE()
: java::sdk::Boolean::FALSE();
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
} else if (aData.isInt32()) {
aOut = java::sdk::Integer::ValueOf(aData.toInt32());
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
} else if (aData.isNumber()) {
aOut = java::sdk::Double::New(aData.toNumber());
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
} else if (aData.isString()) {
return BoxString(aCx, aData, aOut);
} else if (aData.isObject()) {
return BoxObject(aCx, aData, aOut);
} else {
NS_WARNING("Unknown type");
return NS_ERROR_INVALID_ARG;
}
return NS_OK;
}
nsresult
BoxData(const nsAString& aEvent, JSContext* aCx, JS::HandleValue aData,
jni::Object::LocalRef& aOut, bool aObjectOnly)
{
nsresult rv = NS_ERROR_INVALID_ARG;
if (!aObjectOnly) {
rv = BoxValue(aCx, aData, aOut);
} else if (aData.isObject() || aData.isNullOrUndefined()) {
rv = BoxObject(aCx, aData, aOut);
}
if (rv != NS_ERROR_INVALID_ARG) {
return rv;
}
NS_ConvertUTF16toUTF8 event(aEvent);
if (JS_IsExceptionPending(aCx)) {
JS_ReportWarningUTF8(aCx, "Error dispatching %s", event.get());
} else {
JS_ReportErrorUTF8(aCx, "Invalid event data for %s", event.get());
}
return NS_ERROR_INVALID_ARG;
}
nsresult
UnboxString(JSContext* aCx, const jni::Object::LocalRef& aData,
JS::MutableHandleValue aOut)
{
if (!aData) {
aOut.setNull();
return NS_OK;
}
MOZ_ASSERT(aData.IsInstanceOf<jni::String>());
JNIEnv* const env = aData.Env();
const jstring jstr = jstring(aData.Get());
const size_t len = env->GetStringLength(jstr);
const jchar* const jchars = env->GetStringChars(jstr, nullptr);
if (NS_WARN_IF(!jchars)) {
env->ExceptionClear();
return NS_ERROR_FAILURE;
}
auto releaseStr = MakeScopeExit([env, jstr, jchars] {
env->ReleaseStringChars(jstr, jchars);
env->ExceptionClear();
});
JS::RootedString str(aCx, JS_NewUCStringCopyN(
aCx, reinterpret_cast<const char16_t*>(jchars), len));
NS_ENSURE_TRUE(CheckJS(aCx, !!str), NS_ERROR_FAILURE);
aOut.setString(str);
return NS_OK;
}
nsresult
UnboxValue(JSContext* aCx, const jni::Object::LocalRef& aData,
JS::MutableHandleValue aOut);
nsresult
UnboxBundle(JSContext* aCx, const jni::Object::LocalRef& aData,
JS::MutableHandleValue aOut)
{
if (!aData) {
aOut.setNull();
return NS_OK;
}
MOZ_ASSERT(aData.IsInstanceOf<java::GeckoBundle>());
JNIEnv* const env = aData.Env();
const auto& bundle = java::GeckoBundle::Ref::From(aData);
jni::ObjectArray::LocalRef keys = bundle->Keys();
jni::ObjectArray::LocalRef values = bundle->Values();
const size_t len = keys->Length();
JS::RootedObject obj(aCx, JS_NewPlainObject(aCx));
NS_ENSURE_TRUE(CheckJS(aCx, !!obj), NS_ERROR_FAILURE);
NS_ENSURE_TRUE(values->Length() == len, NS_ERROR_FAILURE);
for (size_t i = 0; i < len; i++) {
jni::String::LocalRef key = keys->GetElement(i);
const size_t keyLen = env->GetStringLength(key.Get());
const jchar* const keyChars = env->GetStringChars(key.Get(), nullptr);
if (NS_WARN_IF(!keyChars)) {
env->ExceptionClear();
return NS_ERROR_FAILURE;
}
auto releaseKeyChars = MakeScopeExit([env, &key, keyChars] {
env->ReleaseStringChars(key.Get(), keyChars);
env->ExceptionClear();
});
JS::RootedValue value(aCx);
nsresult rv = UnboxValue(aCx, values->GetElement(i), &value);
if (rv == NS_ERROR_INVALID_ARG && !JS_IsExceptionPending(aCx)) {
JS_ReportErrorUTF8(
aCx, u8"Invalid event data property %s",
NS_ConvertUTF16toUTF8(nsString(reinterpret_cast<
const char16_t*>(keyChars), keyLen)).get());
}
NS_ENSURE_SUCCESS(rv, rv);
NS_ENSURE_TRUE(CheckJS(aCx, JS_SetUCProperty(aCx, obj, reinterpret_cast<
const char16_t*>(keyChars), keyLen, value)),
NS_ERROR_FAILURE);
}
aOut.setObject(*obj);
return NS_OK;
}
template<typename Type, typename JNIType, typename ArrayType,
JNIType* (JNIEnv::*GetElements)(ArrayType, jboolean*),
void (JNIEnv::*ReleaseElements)(ArrayType, JNIType*, jint),
JS::Value (*ToValue)(Type)> nsresult
UnboxArrayPrimitive(JSContext* aCx, const jni::Object::LocalRef& aData,
JS::MutableHandleValue aOut)
{
JNIEnv* const env = aData.Env();
const ArrayType jarray = ArrayType(aData.Get());
JNIType* const array = (env->*GetElements)(jarray, nullptr);
JS::AutoValueVector elements(aCx);
if (NS_WARN_IF(!array)) {
env->ExceptionClear();
return NS_ERROR_FAILURE;
}
auto releaseArray = MakeScopeExit([env, jarray, array] {
(env->*ReleaseElements)(jarray, array, JNI_ABORT);
env->ExceptionClear();
});
const size_t len = env->GetArrayLength(jarray);
NS_ENSURE_TRUE(elements.initCapacity(len), NS_ERROR_FAILURE);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
for (size_t i = 0; i < len; i++) {
NS_ENSURE_TRUE(elements.append((*ToValue)(Type(array[i]))),
NS_ERROR_FAILURE);
}
JS::RootedObject obj(aCx, JS_NewArrayObject(
aCx, JS::HandleValueArray(elements)));
NS_ENSURE_TRUE(CheckJS(aCx, !!obj), NS_ERROR_FAILURE);
aOut.setObject(*obj);
return NS_OK;
}
struct StringArray : jni::ObjectBase<StringArray>
{
static const char name[];
};
struct GeckoBundleArray : jni::ObjectBase<GeckoBundleArray>
{
static const char name[];
};
const char StringArray::name[] = "[Ljava/lang/String;";
const char GeckoBundleArray::name[] = "[Lorg/mozilla/gecko/util/GeckoBundle;";
template<nsresult (*Unbox)(JSContext*, const jni::Object::LocalRef&,
JS::MutableHandleValue)> nsresult
UnboxArrayObject(JSContext* aCx, const jni::Object::LocalRef& aData,
JS::MutableHandleValue aOut)
{
jni::ObjectArray::LocalRef array(
aData.Env(), jni::ObjectArray::Ref::From(aData));
const size_t len = array->Length();
JS::RootedObject obj(aCx, JS_NewArrayObject(aCx, len));
NS_ENSURE_TRUE(CheckJS(aCx, !!obj), NS_ERROR_FAILURE);
for (size_t i = 0; i < len; i++) {
jni::Object::LocalRef element = array->GetElement(i);
JS::RootedValue value(aCx);
nsresult rv = (*Unbox)(aCx, element, &value);
NS_ENSURE_SUCCESS(rv, rv);
NS_ENSURE_TRUE(CheckJS(aCx, JS_SetElement(aCx, obj, i, value)),
NS_ERROR_FAILURE);
}
aOut.setObject(*obj);
return NS_OK;
}
template<class T> jfieldID
GetValueFieldID(const char* aType)
{
MOZ_ASSERT(NS_IsMainThread());
JNIEnv* const env = jni::GetGeckoThreadEnv();
const jfieldID id = env->GetFieldID(
typename T::Context(env, nullptr).ClassRef(), "value", aType);
env->ExceptionClear();
return id;
}
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
nsresult
UnboxValue(JSContext* aCx, const jni::Object::LocalRef& aData,
JS::MutableHandleValue aOut)
{
static jfieldID booleanValueField =
GetValueFieldID<java::sdk::Boolean>("Z");
static jfieldID intValueField = GetValueFieldID<java::sdk::Integer>("I");
static jfieldID doubleValueField = GetValueFieldID<java::sdk::Double>("D");
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
if (!aData) {
aOut.setNull();
} else if (aData.IsInstanceOf<jni::Boolean>()) {
if (booleanValueField) {
aOut.setBoolean(aData.Env()->GetBooleanField(
aData.Get(), booleanValueField) != JNI_FALSE);
MOZ_CATCH_JNI_EXCEPTION(aData.Env());
} else {
aOut.setBoolean(
java::sdk::Boolean::Ref::From(aData)->BooleanValue());
}
} else if (aData.IsInstanceOf<jni::Integer>()) {
if (intValueField) {
aOut.setInt32(aData.Env()->GetIntField(aData.Get(), intValueField));
MOZ_CATCH_JNI_EXCEPTION(aData.Env());
} else {
aOut.setInt32(java::sdk::Number::Ref::From(aData)->IntValue());
}
} else if (aData.IsInstanceOf<jni::Byte>() ||
aData.IsInstanceOf<jni::Short>()) {
aOut.setInt32(java::sdk::Number::Ref::From(aData)->IntValue());
} else if (aData.IsInstanceOf<jni::Double>()) {
if (doubleValueField) {
aOut.setNumber(aData.Env()->GetDoubleField(
aData.Get(), doubleValueField));
} else {
aOut.setNumber(java::sdk::Number::Ref::From(aData)->DoubleValue());
}
} else if (aData.IsInstanceOf<jni::Float>() ||
aData.IsInstanceOf<jni::Long>()) {
aOut.setNumber(java::sdk::Number::Ref::From(aData)->DoubleValue());
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
} else if (aData.IsInstanceOf<jni::String>()) {
return UnboxString(aCx, aData, aOut);
} else if (aData.IsInstanceOf<jni::Character>()) {
return UnboxString(aCx, java::sdk::String::ValueOf(aData), aOut);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
} else if (aData.IsInstanceOf<java::GeckoBundle>()) {
return UnboxBundle(aCx, aData, aOut);
} else if (aData.IsInstanceOf<jni::BooleanArray>()) {
return UnboxArrayPrimitive<bool, jboolean, jbooleanArray,
&JNIEnv::GetBooleanArrayElements,
&JNIEnv::ReleaseBooleanArrayElements,
&JS::BooleanValue>(aCx, aData, aOut);
} else if (aData.IsInstanceOf<jni::IntArray>()) {
return UnboxArrayPrimitive<int32_t, jint, jintArray,
&JNIEnv::GetIntArrayElements,
&JNIEnv::ReleaseIntArrayElements,
&JS::Int32Value>(aCx, aData, aOut);
} else if (aData.IsInstanceOf<jni::DoubleArray>()) {
return UnboxArrayPrimitive<double, jdouble, jdoubleArray,
&JNIEnv::GetDoubleArrayElements,
&JNIEnv::ReleaseDoubleArrayElements,
&JS::DoubleValue>(aCx, aData, aOut);
} else if (aData.IsInstanceOf<StringArray>()) {
return UnboxArrayObject<&UnboxString>(aCx, aData, aOut);
} else if (aData.IsInstanceOf<GeckoBundleArray>()) {
return UnboxArrayObject<&UnboxBundle>(aCx, aData, aOut);
} else {
NS_WARNING("Invalid type");
return NS_ERROR_INVALID_ARG;
}
return NS_OK;
}
nsresult
UnboxData(jni::String::Param aEvent, JSContext* aCx, jni::Object::Param aData,
JS::MutableHandleValue aOut, bool aBundleOnly)
{
MOZ_ASSERT(NS_IsMainThread());
jni::Object::LocalRef jniData(jni::GetGeckoThreadEnv(), aData);
nsresult rv = NS_ERROR_INVALID_ARG;
if (!aBundleOnly) {
rv = UnboxValue(aCx, jniData, aOut);
} else if (!jniData || jniData.IsInstanceOf<java::GeckoBundle>()) {
rv = UnboxBundle(aCx, jniData, aOut);
}
if (rv != NS_ERROR_INVALID_ARG || !aEvent) {
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
return rv;
}
nsCString event = aEvent->ToCString();
if (JS_IsExceptionPending(aCx)) {
JS_ReportWarningUTF8(aCx, "Error dispatching %s", event.get());
} else {
JS_ReportErrorUTF8(aCx, "Invalid event data for %s", event.get());
}
return NS_ERROR_INVALID_ARG;
}
class JavaCallbackDelegate final : public nsIAndroidEventCallback
{
const java::EventCallback::GlobalRef mCallback;
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
virtual ~JavaCallbackDelegate() {}
NS_IMETHOD Call(JS::HandleValue aData,
void (java::EventCallback::*aCall)(
jni::Object::Param) const)
{
MOZ_ASSERT(NS_IsMainThread());
AutoJSContext cx;
jni::Object::LocalRef data(jni::GetGeckoThreadEnv());
nsresult rv = BoxData(NS_LITERAL_STRING("callback"), cx, aData, data,
/* ObjectOnly */ false);
NS_ENSURE_SUCCESS(rv, JS_IsExceptionPending(cx) ? NS_OK : rv);
dom::AutoNoJSAPI nojsapi;
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
(java::EventCallback(*mCallback).*aCall)(data);
return NS_OK;
}
public:
explicit JavaCallbackDelegate(java::EventCallback::Param aCallback)
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
: mCallback(jni::GetGeckoThreadEnv(), aCallback)
{}
NS_DECL_ISUPPORTS
NS_IMETHOD OnSuccess(JS::HandleValue aData) override
{
return Call(aData, &java::EventCallback::SendSuccess);
}
NS_IMETHOD OnError(JS::HandleValue aData) override
{
return Call(aData, &java::EventCallback::SendError);
}
};
NS_IMPL_ISUPPORTS(JavaCallbackDelegate, nsIAndroidEventCallback)
class NativeCallbackDelegateSupport final :
public java::EventDispatcher::NativeCallbackDelegate
::Natives<NativeCallbackDelegateSupport>
{
using CallbackDelegate = java::EventDispatcher::NativeCallbackDelegate;
using Base = CallbackDelegate::Natives<NativeCallbackDelegateSupport>;
const nsCOMPtr<nsIAndroidEventCallback> mCallback;
const nsCOMPtr<nsIAndroidEventFinalizer> mFinalizer;
const nsCOMPtr<nsIGlobalObject> mGlobalObject;
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
void Call(jni::Object::Param aData,
nsresult (nsIAndroidEventCallback::*aCall)(JS::HandleValue))
{
MOZ_ASSERT(NS_IsMainThread());
// Use either the attached window's realm or a default realm.
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
dom::AutoJSAPI jsapi;
NS_ENSURE_TRUE_VOID(jsapi.Init(mGlobalObject));
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
JS::RootedValue data(jsapi.cx());
nsresult rv = UnboxData(NS_LITERAL_STRING("callback"), jsapi.cx(),
aData, &data, /* BundleOnly */ false);
NS_ENSURE_SUCCESS_VOID(rv);
dom::AutoNoJSAPI nojsapi;
rv = (mCallback->*aCall)(data);
NS_ENSURE_SUCCESS_VOID(rv);
}
public:
using Base::AttachNative;
template<typename Functor>
static void OnNativeCall(Functor&& aCall)
{
if (NS_IsMainThread()) {
// Invoke callbacks synchronously if we're already on Gecko thread.
return aCall();
}
NS_DispatchToMainThread(NS_NewRunnableFunction("OnNativeCall", std::move(aCall)));
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
}
static void Finalize(const CallbackDelegate::LocalRef& aInstance)
{
DisposeNative(aInstance);
}
NativeCallbackDelegateSupport(nsIAndroidEventCallback* callback,
nsIAndroidEventFinalizer* finalizer,
nsIGlobalObject* globalObject)
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
: mCallback(callback)
, mFinalizer(finalizer)
, mGlobalObject(globalObject)
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
{}
~NativeCallbackDelegateSupport()
{
if (mFinalizer) {
mFinalizer->OnFinalize();
}
}
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
void SendSuccess(jni::Object::Param aData)
{
Call(aData, &nsIAndroidEventCallback::OnSuccess);
}
void SendError(jni::Object::Param aData)
{
Call(aData, &nsIAndroidEventCallback::OnError);
}
};
class FinalizingCallbackDelegate final : public nsIAndroidEventCallback
{
const nsCOMPtr<nsIAndroidEventCallback> mCallback;
const nsCOMPtr<nsIAndroidEventFinalizer> mFinalizer;
virtual ~FinalizingCallbackDelegate()
{
if (mFinalizer) {
mFinalizer->OnFinalize();
}
}
public:
FinalizingCallbackDelegate(nsIAndroidEventCallback* aCallback,
nsIAndroidEventFinalizer* aFinalizer)
: mCallback(aCallback)
, mFinalizer(aFinalizer)
{}
NS_DECL_ISUPPORTS
NS_FORWARD_NSIANDROIDEVENTCALLBACK(mCallback->);
};
NS_IMPL_ISUPPORTS(FinalizingCallbackDelegate, nsIAndroidEventCallback)
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
} // namespace detail
using namespace detail;
NS_IMPL_ISUPPORTS(EventDispatcher, nsIAndroidEventDispatcher)
nsIGlobalObject*
EventDispatcher::GetGlobalObject()
{
if (mDOMWindow) {
return nsGlobalWindowInner::Cast(mDOMWindow->GetCurrentInnerWindow());
}
return xpc::NativeGlobal(xpc::PrivilegedJunkScope());
}
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
nsresult
EventDispatcher::DispatchOnGecko(ListenersList* list, const nsAString& aEvent,
JS::HandleValue aData,
nsIAndroidEventCallback* aCallback)
{
MOZ_ASSERT(NS_IsMainThread());
dom::AutoNoJSAPI nojsapi;
list->lockCount++;
auto iteratingScope = MakeScopeExit([list] {
list->lockCount--;
if (list->lockCount || !list->unregistering) {
return;
}
list->unregistering = false;
for (ssize_t i = list->listeners.Count() - 1; i >= 0; i--) {
if (list->listeners[i]) {
continue;
}
list->listeners.RemoveObjectAt(i);
}
});
const size_t count = list->listeners.Count();
for (size_t i = 0; i < count; i++) {
if (!list->listeners[i]) {
// Unregistered.
continue;
}
const nsresult rv = list->listeners[i]->OnEvent(
aEvent, aData, aCallback);
Bug 1337467 - Convert observers to bundle events; r=rbarker r=sebastian Bug 1337467 - 1. Convert "Window:Resize" observer to event; r=rbarker Bug 1337467 - 2. Convert "ScrollTo:FocusedInput" observer to event; r=rbarker Bug 1337467 - 3. Convert "Update:CheckResult" observer to event; r=sebastian Also remove notifyCheckUpdateResult from GeckoInterface. Bug 1337467 - 4. Convert "GeckoView:ImportScript" observer to event; r=sebastian Bug 1337467 - 5. Convert accessibility observers to events; r=sebastian Bug 1337467 - 6. Convert media/casting observers to events; r=sebastian Bug 1337467 - 7. Convert "Sanitize:ClearData" observer to event; r=sebastian Bug 1337467 - 8. Convert "Notification:Event" observer to event; r=sebastian Bug 1337467 - 9. Convert BrowserApp observers to events; r=sebastian Bug 1337467 - 10. Convert Tab observers to events; r=sebastian Bug 1337467 - 11. Convert "Passwords:Init" and "FormHistory:Init" observers to events; r=sebastian Bug 1337467 - 12. Convert Reader observers to events; r=sebastian Bug 1337467 - 13. Convert Distribution observers to events; r=sebastian Bug 1337467 - 14. Convert "Fonts:Reload" observer to event; r=sebastian Bug 1337467 - 15. Convert RecentTabsAdapter observers to events; r=sebastian Bug 1337467 - 16. Convert "Session:Prefetch" observer to event; r=sebastian Bug 1337467 - 17. Convert "Browser:Quit" and "FullScreen:Exit" observers to events; r=sebastian Bug 1337467 - 18. Convert SessionStore observers to events; r=sebastian The "Session:NotifyLocationChange" observer is sent by browser.js and requires passing a browser reference, so it's left as an observer. Bug 1337467 - 19. Remove unused "Tab:Screenshot:Cancel" notifyObserver call; r=me Bug 1337467 - 20. Convert "Session:Navigate" observer to event; r=sebastian Bug 1337467 - 21. Convert "Locale:*" observers to events; r=sebastian Bug 1337467 - Add log for unhandled events; r=me Add back the log indicating no listener for an event, which can be useful when reading logcat. r=me for trivial change. Bug 1337467 - Don't return error from EventDispatcher when OnEvent fails; r=me When a listener's OnEvent method returns an error, continue to dispatch to other listeners and don't return an error from the dispatch function. This avoids unexpected errors when dispatching events. r=me for trivial patch.
2017-03-07 20:34:04 +03:00
Unused << NS_WARN_IF(NS_FAILED(rv));
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
}
return NS_OK;
}
java::EventDispatcher::NativeCallbackDelegate::LocalRef
EventDispatcher::WrapCallback(nsIAndroidEventCallback* aCallback,
nsIAndroidEventFinalizer* aFinalizer)
{
if (!aCallback) {
return java::EventDispatcher::NativeCallbackDelegate::LocalRef(
jni::GetGeckoThreadEnv());
}
java::EventDispatcher::NativeCallbackDelegate::LocalRef
callback = java::EventDispatcher::NativeCallbackDelegate::New();
NativeCallbackDelegateSupport::AttachNative(
callback,
MakeUnique<NativeCallbackDelegateSupport>(
aCallback, aFinalizer, GetGlobalObject()));
return callback;
}
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
NS_IMETHODIMP
EventDispatcher::Dispatch(JS::HandleValue aEvent, JS::HandleValue aData,
nsIAndroidEventCallback* aCallback,
nsIAndroidEventFinalizer* aFinalizer, JSContext* aCx)
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
{
MOZ_ASSERT(NS_IsMainThread());
if (!aEvent.isString()) {
NS_WARNING("Invalid event name");
return NS_ERROR_INVALID_ARG;
}
nsAutoJSString event;
NS_ENSURE_TRUE(CheckJS(aCx, event.init(aCx, aEvent.toString())),
NS_ERROR_OUT_OF_MEMORY);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
// Don't need to lock here because we're on the main thread, and we can't
// race against Register/UnregisterListener.
ListenersList* list = mListenersMap.Get(event);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
if (list) {
if (!aCallback || !aFinalizer) {
return DispatchOnGecko(list, event, aData, aCallback);
}
nsCOMPtr<nsIAndroidEventCallback> callback(
new FinalizingCallbackDelegate(aCallback, aFinalizer));
return DispatchOnGecko(list, event, aData, callback);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
}
java::EventDispatcher::LocalRef dispatcher(mDispatcher);
if (!dispatcher) {
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
return NS_OK;
}
jni::Object::LocalRef data(jni::GetGeckoThreadEnv());
nsresult rv = BoxData(event, aCx, aData, data, /* ObjectOnly */ true);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
NS_ENSURE_SUCCESS(rv, JS_IsExceptionPending(aCx) ? NS_OK : rv);
dom::AutoNoJSAPI nojsapi;
dispatcher->DispatchToThreads(event, data,
WrapCallback(aCallback, aFinalizer));
return NS_OK;
}
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
nsresult
EventDispatcher::Dispatch(const char16_t* aEvent,
java::GeckoBundle::Param aData,
nsIAndroidEventCallback* aCallback)
{
nsDependentString event(aEvent);
ListenersList* list = mListenersMap.Get(event);
if (list) {
dom::AutoJSAPI jsapi;
NS_ENSURE_TRUE(jsapi.Init(GetGlobalObject()),
NS_ERROR_FAILURE);
JS::RootedValue data(jsapi.cx());
nsresult rv = UnboxData(/* Event */ nullptr, jsapi.cx(), aData, &data,
/* BundleOnly */ true);
NS_ENSURE_SUCCESS(rv, rv);
return DispatchOnGecko(list, event, data, aCallback);
}
java::EventDispatcher::LocalRef dispatcher(mDispatcher);
if (!dispatcher) {
return NS_OK;
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
}
dispatcher->DispatchToThreads(event, aData, WrapCallback(aCallback));
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
return NS_OK;
}
nsresult
EventDispatcher::IterateEvents(JSContext* aCx, JS::HandleValue aEvents,
IterateEventsCallback aCallback,
nsIAndroidEventListener* aListener)
{
MOZ_ASSERT(NS_IsMainThread());
MutexAutoLock lock(mLock);
auto processEvent = [this, aCx, aCallback, aListener]
(JS::HandleValue event) -> nsresult {
nsAutoJSString str;
NS_ENSURE_TRUE(CheckJS(aCx, str.init(aCx, event.toString())),
NS_ERROR_OUT_OF_MEMORY);
return (this->*aCallback)(str, aListener);
};
if (aEvents.isString()) {
return processEvent(aEvents);
}
bool isArray = false;
NS_ENSURE_TRUE(aEvents.isObject(), NS_ERROR_INVALID_ARG);
NS_ENSURE_TRUE(CheckJS(aCx, JS_IsArrayObject(aCx, aEvents, &isArray)),
NS_ERROR_INVALID_ARG);
NS_ENSURE_TRUE(isArray, NS_ERROR_INVALID_ARG);
JS::RootedObject events(aCx, &aEvents.toObject());
uint32_t length = 0;
NS_ENSURE_TRUE(CheckJS(aCx, JS_GetArrayLength(aCx, events, &length)),
NS_ERROR_INVALID_ARG);
NS_ENSURE_TRUE(length, NS_ERROR_INVALID_ARG);
for (size_t i = 0; i < length; i++) {
JS::RootedValue event(aCx);
NS_ENSURE_TRUE(CheckJS(aCx, JS_GetElement(aCx, events, i, &event)),
NS_ERROR_INVALID_ARG);
NS_ENSURE_TRUE(event.isString(), NS_ERROR_INVALID_ARG);
const nsresult rv = processEvent(event);
NS_ENSURE_SUCCESS(rv, rv);
}
return NS_OK;
}
nsresult
EventDispatcher::RegisterEventLocked(const nsAString& aEvent,
nsIAndroidEventListener* aListener)
{
ListenersList* list = mListenersMap.Get(aEvent);
if (!list) {
list = new ListenersList();
mListenersMap.Put(aEvent, list);
}
#ifdef DEBUG
for (ssize_t i = 0; i < list->listeners.Count(); i++) {
NS_ENSURE_TRUE(list->listeners[i] != aListener,
NS_ERROR_ALREADY_INITIALIZED);
}
#endif
list->listeners.AppendObject(aListener);
return NS_OK;
}
NS_IMETHODIMP
EventDispatcher::RegisterListener(nsIAndroidEventListener* aListener,
JS::HandleValue aEvents, JSContext *aCx)
{
return IterateEvents(aCx, aEvents, &EventDispatcher::RegisterEventLocked,
aListener);
}
nsresult
EventDispatcher::UnregisterEventLocked(const nsAString& aEvent,
nsIAndroidEventListener* aListener)
{
ListenersList* list = mListenersMap.Get(aEvent);
#ifdef DEBUG
NS_ENSURE_TRUE(list, NS_ERROR_NOT_INITIALIZED);
#else
NS_ENSURE_TRUE(list, NS_OK);
#endif
DebugOnly<bool> found = false;
for (ssize_t i = list->listeners.Count() - 1; i >= 0; i--) {
if (list->listeners[i] != aListener) {
continue;
}
if (list->lockCount) {
// Only mark for removal when list is locked.
list->listeners.ReplaceObjectAt(nullptr, i);
list->unregistering = true;
} else {
list->listeners.RemoveObjectAt(i);
}
found = true;
}
#ifdef DEBUG
return found ? NS_OK : NS_ERROR_NOT_INITIALIZED;
#else
return NS_OK;
#endif
}
NS_IMETHODIMP
EventDispatcher::UnregisterListener(nsIAndroidEventListener* aListener,
JS::HandleValue aEvents, JSContext *aCx)
{
return IterateEvents(aCx, aEvents, &EventDispatcher::UnregisterEventLocked,
aListener);
}
void
EventDispatcher::Attach(java::EventDispatcher::Param aDispatcher,
nsPIDOMWindowOuter* aDOMWindow)
{
MOZ_ASSERT(NS_IsMainThread());
MOZ_ASSERT(aDispatcher);
java::EventDispatcher::LocalRef dispatcher(mDispatcher);
if (dispatcher) {
if (dispatcher == aDispatcher) {
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
// Only need to update the window.
mDOMWindow = aDOMWindow;
return;
}
dispatcher->SetAttachedToGecko(java::EventDispatcher::REATTACHING);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
}
dispatcher = java::EventDispatcher::LocalRef(aDispatcher);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
NativesBase::AttachNative(dispatcher, this);
mDispatcher = dispatcher;
mDOMWindow = aDOMWindow;
dispatcher->SetAttachedToGecko(java::EventDispatcher::ATTACHED);
}
void
EventDispatcher::Detach()
{
MOZ_ASSERT(NS_IsMainThread());
MOZ_ASSERT(mDispatcher);
java::EventDispatcher::GlobalRef dispatcher(mDispatcher);
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
// SetAttachedToGecko will call disposeNative for us later on the Gecko
// thread to make sure all pending dispatchToGecko calls have completed.
if (dispatcher) {
dispatcher->SetAttachedToGecko(java::EventDispatcher::DETACHED);
}
mDispatcher = nullptr;
mDOMWindow = nullptr;
}
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
bool
EventDispatcher::HasGeckoListener(jni::String::Param aEvent)
{
// Can be called from any thread.
MutexAutoLock lock(mLock);
return !!mListenersMap.Get(aEvent->ToString());
}
void
EventDispatcher::DispatchToGecko(jni::String::Param aEvent,
jni::Object::Param aData,
jni::Object::Param aCallback)
{
MOZ_ASSERT(NS_IsMainThread());
// Don't need to lock here because we're on the main thread, and we can't
// race against Register/UnregisterListener.
nsString event = aEvent->ToString();
ListenersList* list = mListenersMap.Get(event);
if (!list || list->listeners.IsEmpty()) {
return;
}
// Use the same compartment as the attached window if possible, otherwise
// use a default compartment.
dom::AutoJSAPI jsapi;
NS_ENSURE_TRUE_VOID(jsapi.Init(GetGlobalObject()));
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
JS::RootedValue data(jsapi.cx());
nsresult rv = UnboxData(aEvent, jsapi.cx(), aData, &data,
/* BundleOnly */ true);
NS_ENSURE_SUCCESS_VOID(rv);
nsCOMPtr<nsIAndroidEventCallback> callback;
if (aCallback) {
callback = new JavaCallbackDelegate(
java::EventCallback::Ref::From(aCallback));
}
DispatchOnGecko(list, event, data, callback);
}
/* static */
nsresult
EventDispatcher::UnboxBundle(JSContext* aCx, jni::Object::Param aData,
JS::MutableHandleValue aOut)
{
return detail::UnboxBundle(aCx, aData, aOut);
}
Bug 1307820 - Implement per-GeckoView messaging; r=snorp r=sebastian Bug 1307820 - 1a. Move GeckoApp EventDispatcher to GeckoView; r=snorp Make it a GeckoView-specific EventDispatcher instead of GeckoApp-specific, so that GeckoView consumers can benefit from a per-view EventDispatcher. In addition, a few events like Gecko:Ready are moved back to the global EventDispatcher because that makes more sense. Bug 1307820 - 1b. Don't use GeckoApp EventDispatcher during inflation; r=snorp During layout inflation, we don't yet have GeckoView and therefore the GeckoView EventDispatcher, so we should not register events until later, typically during onAttachedToWindow. Bug 1307820 - 2. Introduce GeckoBundle; r=snorp The Android Bundle class has several disadvantages when used for holding structured data from JS. The most obvious one is the differentiation between int and double, which doesn't exist in JS. So when a JS number is converted to either a Bundle int or double, we run the risk of making a wrong conversion, resulting in a type mismatch exception when Java uses the Bundle. This extends to number arrays from JS. There is one more gotcha when using arrays. When we receive an empty array from JS, there is no way for us to determine the type of the array, because even empty arrays in Java have types. We are forced to pick an arbitrary type like boolean[], which can easily result in a type mismatch exception when using the array on the Java side. In addition, Bundle is fairly cumbersome, and we cannot access the inner structures of Bundle from Java or JNI, making it harder to use. With these factors in mind, this patch introduces GeckoBundle as a better choice for Gecko/Java communication. It is almost fully API-compatible with the Android Bundle; only the Bundle array methods are different. It resolves the numbers problem by performing conversions if necessary, and it is a lot more lightweight than Bundle. Bug 1307820 - 3. Convert BundleEventListener to use GeckoBundle; r=snorp Convert BundleEventListener from using Bundle to using GeckoBundle. Because NativeJSContainer still only supports Bundle, we do an extra conversion when sending Bundle messages, but eventually, as we eliminate the use of NativeJSContainer, that will go away as well. Bug 1307820 - 4. Introduce EventDispatcher interfaces; r=snorp Introduce several new XPCOM interfaces for the new EventDispatcher API, these interfaces are mostly mirrored after their Java counterparts. * nsIAndroidEventDispatcher is the main interface for registering/unregistering listeners and for dispatching events from JS/C++. * nsIAndroidEventListener is the interface that JS/C++ clients implement to receive events. * nsIAndroidEventCallback is the interface that JS/C++ clients implement to receive responses from dispatched events. * nsIAndroidView is the new interface that every window receives that is specific to the window/GeckoView pair. It is passed to chrome scripts through window arguments. Bug 1307820 - 5. Remove EventDispatcher references from gfx code; r=snorp EventDispatcher was used for JPZC, but NPZC doesn't use it anymore. Bug 1307820 - 6. General JNI template improvements; r=snorp This patch includes several improvements to the JNI templates. * Context::RawClassRef is removed to avoid misuse, as Context::ClassRef should be used instead. * Fix a compile error, in certain usages, in the DisposeNative overload in NativeStub. * Add Ref::IsInstanceOf and Context::IsInstanceOf to mirror the JNIEnv::IsInstanceOf call. * Add Ref::operator* and Context::operator* to provide an easy way to get a Context object. * Add built-in declarations for boxed Java objects (e.g. Boolean, Integer, etc). * Add ObjectArray::New for creating new object arrays of specific types. * Add lvalue qualifiers to LocalRef::operator= and GlobalRef::operator=, to prevent accidentally assigning to rvalues. (e.g. `objectArray->GetElement(0) = newObject;`, which won't work as intended.) Bug 1307820 - 7. Support ownership through RefPtr for native JNI objects; r=snorp In addition to direct ownership and weak pointer ownership, add a third ownership model where a native JNI object owns a RefPtr that holds a strong reference to the actual C++ object. This ownership model works well with ref-counted objects such as XPCOM objects, and is activated through the presence of public members AddRef() and Release() in the C++ object. Bug 1307820 - 8. Implement Gecko-side EventDispatcher; r=snorp Add a skeletal implementation of EventDispatcher on the Gecko side. Each widget::EventDispatcher will be associated with a Java EventDispatcher, so events can be dispatched from Gecko to Java and vice versa. AndroidBridge and nsWindow will implement nsIAndroidEventDispatcher through widget::EventDispatcher. Other patches will add more complete functionality such as GeckoBundle/JSObject translation and support for callbacks. Bug 1307820 - 9. Implement dispatching between Gecko/Java; r=snorp Implement translation between JSObject and GeckoBundle, and use that for dispatching events from Gecko to Java and vice versa. Bug 1307820 - 10. Implement callback support; r=snorp Implement callback support for both Gecko-to-Java events and Java-to-Gecko events. For Gecko-to-Java, we translate nsIAndroidEventCallback to a Java EventCallback through NativeCallbackDelegate and pass it to the Java listener. For Java-to-Gecko, we translate EventCallback to a nsIAndroidEventCallback through JavaCallbackDelegate and pass it to the Gecko listener. There is another JavaCallbackDelegate on the Java side that redirects the callback to a particular thread. For example, if the event was dispatched from the UI thread, we make sure the callback happens on the UI thread as well. Bug 1307820 - 11. Add BundleEventListener support for Gecko thread; r=snorp Add support for BundleEventListener on the Gecko thread, so that we can use it to replace any existing GeckoEventListener or NativeEventListener implementations that require the listener be run synchronously on the Gecko thread. Bug 1307820 - 12. Add global EventDispatcher in AndroidBridge; r=snorp Add an instance of EventDispatcher to AndroidBridge to act as a global event dispatcher. Bug 1307820 - 13. Add per-nsWindow EventDispatcher; r=snorp Add an instance of EventDispatcher to each nsWindow through an AndroidView object, which implements nsIAndroidView. The nsIAndroidView is passed to the chrome script through the window argument when opening the window. Bug 1307820 - 14. Update auto-generated bindings; r=me Bug 1307820 - 15. Update testEventDispatcher; r=snorp Update testEventDispatcher to include new functionalities in EventDisptcher. * Add tests for dispatching events to UI/background thread through nsIAndroidEventDispatcher::dispatch. * Add tests for dispatching events to UI/background thread through EventDispatcher.dispatch. * Add tests for dispatching events to Gecko thread through EventDispatcher.dispatch. Each kind of test exercises both the global EventDispatcher through EventDispatcher.getInstance() and the per-GeckoView EventDispatcher through GeckoApp.getEventDispatcher().
2016-11-14 16:29:50 +03:00
} // namespace widget
} // namespace mozilla