转载

RN 通信

JS 桥

Android: Webkit 的 JavaScriptCore ios: 自带的 javascriptcore

在 Android 的代码,其提供了一个 CatalystInstance 的接口,来做 JS 与 Native 的高度抽象的接口:

public interface CatalystInstance
    extends MemoryPressureListener, JSInstance {
  void runJSBundle();

  // Returns the status of running the JS bundle; waits for an answer if runJSBundle is running
  boolean hasRunJSBundle();

  /**
   * Return the source URL of the JS Bundle that was run, or {@code null} if no JS
   * bundle has been run yet.
   */
  @Nullable String getSourceURL();

  // This is called from java code, so it won't be stripped anyway, but proguard will rename it,
  // which this prevents.
  @Override @DoNotStrip
  void invokeCallback(
      int callbackID,
      NativeArray arguments);
  @DoNotStrip
  void callFunction(
      String module,
      String method,
      NativeArray arguments);
  /**
   * Destroys this catalyst instance, waiting for any other threads in ReactQueueConfiguration
   * (besides the UI thread) to finish running. Must be called from the UI thread so that we can
   * fully shut down other threads.
   */
  void destroy();
  boolean isDestroyed();

  /**
   * Initialize all the native modules
   */
  @VisibleForTesting
  void initialize();

  ReactQueueConfiguration getReactQueueConfiguration();

  <T extends JavaScriptModule> T getJSModule(Class<T> jsInterface);
  <T extends NativeModule> boolean hasNativeModule(Class<T> nativeModuleInterface);
  <T extends NativeModule> T getNativeModule(Class<T> nativeModuleInterface);
  <T extends JSIModule> T getJSIModule(Class<T> jsiModuleInterface);
  Collection<NativeModule> getNativeModules();

  /**
   * This method permits a CatalystInstance to extend the known
   * Native modules. This provided registry contains only the new modules to load.
   */
  void extendNativeModules(NativeModuleRegistry modules);

  /**
   * Adds a idle listener for this Catalyst instance. The listener will receive notifications
   * whenever the bridge transitions from idle to busy and vice-versa, where the busy state is
   * defined as there being some non-zero number of calls to JS that haven't resolved via a
   * onBatchCompleted call. The listener should be purely passive and not affect application logic.
   */
  void addBridgeIdleDebugListener(NotThreadSafeBridgeIdleDebugListener listener);

  /**
   * Removes a NotThreadSafeBridgeIdleDebugListener previously added with
   * {@link #addBridgeIdleDebugListener}
   */
  void removeBridgeIdleDebugListener(NotThreadSafeBridgeIdleDebugListener listener);

  /** This method registers the file path of an additional JS segment by its ID. */
  void registerSegment(int segmentId, String path);

  @VisibleForTesting
  void setGlobalVariable(String propName, String jsonValue);

  /**
   * Get the C pointer (as a long) to the JavaScriptCore context associated with this instance.
   *
   * <p>Use the following pattern to ensure that the JS context is not cleared while you are using
   * it: JavaScriptContextHolder jsContext = reactContext.getJavaScriptContextHolder()
   * synchronized(jsContext) { nativeThingNeedingJsContext(jsContext.get()); }
   */
  JavaScriptContextHolder getJavaScriptContextHolder();

  void addJSIModules(List<JSIModuleHolder> jsiModules);
}

其中主要有 NativeModule 和 JSModule 获取的方法,并调用 JS 侧方法 invokeCallback 和 callFunction 方法。

执行流程

1.构建 ReactNativeHost

在 ReactNativeHost 中的 getPackages 方法中,我们需要传递实现的 ReactPackage 接口,接口中则定义了实现的 NativeModule 集合和 ViewManager 的集合。

2.创建 ReactInstanceManger

ReactNativeHost 是在 application 中提供获取的,而通过 ReactNativeHost 可以获取到 ReactInstanceManager 类,在创建 ReactInstanceManager 的过程中,会得到所有的 ReactPackage,将其保存在 ReactInstanceManager 类中。

这里的 ReactPackage 的获取,会优先添加 CoreModulesPackage,其提供了 Android 基础的 Module功能,如 AndroidInfoModule、DeviceEventManageModule、UIManagerModule等。

这里的 UIManagerModule,则是来负责 UI 视图创建的 View 集合,并接受来自 View 的命令来操作更新 View。

3.ReactRootView 执行 startReactApplication

在一个 RN 的 activity 中,其会在 onCreate 方法中,执行 ReactRootView 的创建工作,之后便是根据创建好的 ReactInstanceManager,来执行 ReactRootView 的 startReactApplication 的方法。

其中会调用 ReactInstanceManager 的 createReactContentInBackground 方法。这里会根据特定的 JS 来源,在一个新的线程执行 createReactContext 方法来创建一个ReactApplicationContext。

4.createReactContext

此方法中,首先会处理之前的 ReactPackages 来生成一个 NativeModuleRegistry,其是用来管理 NativeModule 并提供获取调用的功能。

紧接着,便是利用 CatalystInstanceImpl Builder 构造一个 CatalystInstance。在 CatalystInstanceImpl 的初始化方法中,其会利用传递过来的参数,调用 JNI 方法 initializeBridge,其会执行桥的初始化操作,如获取 NativeModule 至 JNI 中的 ModuleRegistry 中。

之后便是调用 CatalystInstance 的 runJSBundle 方法。其会执行 JSBundleLoader 的 loadScript 方法,这里又会调用回 CatalystInstanceImpl 中的几个提供的相关从 jni 中 load JS 内容的方法。这里主要来完成 JS 内容的加载。

5.setupReactContext

之后,在 NativeModulesQueueThread 上执行 setupReactContext 的方法,这里会执行 CatalyInstance 的 initalize,主要内容则是初始化所有的 NativeModule。

随之,便是调用 attachRootViewToInstance 方法,其参数为 ReactRootView 和 CatalystInstance。其中最主要的内容便是调用 ReactRootView 的 invokeJsEntryPoint :

  /*package */ void invokeJSEntryPoint() {
    if (mJSEntryPoint == null) {
      defaultJSEntryPoint();
    } else {
      mJSEntryPoint.run();
    }
  }

其中 defaultJsEntryPoint 方法,则是调用

  /**
   * Calls the default entry point into JS which is AppRegistry.runApplication()
   */
  private void defaultJSEntryPoint() {
      Systrace.beginSection(TRACE_TAG_REACT_JAVA_BRIDGE, "ReactRootView.runApplication");
      try {
        if (mReactInstanceManager == null || !mIsAttachedToInstance) {
          return;
        }

        ReactContext reactContext = mReactInstanceManager.getCurrentReactContext();
        if (reactContext == null) {
          return;
        }

        CatalystInstance catalystInstance = reactContext.getCatalystInstance();

        WritableNativeMap appParams = new WritableNativeMap();
        appParams.putDouble("rootTag", getRootViewTag());
        @Nullable Bundle appProperties = getAppProperties();
        if (appProperties != null) {
          appParams.putMap("initialProps", Arguments.fromBundle(appProperties));
        }
        if (getUIManagerType() == FABRIC) {
          appParams.putBoolean("fabric", true);
        }

        mShouldLogContentAppeared = true;

        String jsAppModuleName = getJSModuleName();
        catalystInstance.getJSModule(AppRegistry.class).runApplication(jsAppModuleName, appParams);
      } finally {
        Systrace.endSection(TRACE_TAG_REACT_JAVA_BRIDGE);
      }
  }

可以看出,这里的关键内容便是将 rootTag、initialProps 参数内容,放置在 appParams 中,调用 JS 侧的对象 AppRegistry,并调用其 runApplication 方法,来完成 JS 端内容的调用。

JS 调用 Native

将 Native 的方法提供给 JS 来调用,

1.步骤

  • 1)Native 端定义实现 NativeModule 接口
  • 2)将此 NativeModule 接口添加至 ReactPackage 中
  • 3)JS 端在 NativeModule.js 中获取到指定的 Module 对象,根据提供的方法进行调用。

2.实现

1.获取 Native 对象

在获取一个 Native 端提供的类、对象,通常系统的,我们都是通过 ‘react-native’ 中获取,而我们自定义的模块则需要从 NativeModule 中获取,以 ToastAndroid 为例:

import {
  ToastAndroid,
} from 'react-native';

其中,ToastAndroid 对应着 ToastAndroid.android.js,其文件中的 ToastAndroid 内容则是 RCTToastAndroid,其获取方式如下:

const { ToastAndroid} = require('NativeModule')
const RCTToastAndroid = require('NativeModules').ToastAndroid;

2.NativeModule 属性对象的赋值

在 NativeModule.js 中,NativeModule 对象的生成过程中,其已经调用了 NativeModule 的 genModule 方法,生成了相应的 Module,并赋值了 NativeModule 对象。

其中 genModule 的参数为 ModuleConfig 和 moduleID 的数字类型值。ModuleConfig 的定义如下:

type ModuleConfig = [
  string, /* name */
  ?Object, /* constants */
  Array<string>, /* functions */
  Array<number>, /* promise method IDs */
  Array<number>, /* sync method IDs */
];

它包含了一个模块的名称、常量值、方法。其中 promise 及 sync 的方法采用方法数组索引来表示。 另外,moduleID 为 module 数组中的索引值。

而调用 genModule 方法,则是根据此 config 信息,生成一个对应的 module 对象:

  const [moduleName, constants, methods, promiseMethods, syncMethods] = config;

  if (!constants && !methods) {
    // Module contents will be filled in lazily later
    return { name: moduleName };
  }

  const module = {};
  methods && methods.forEach((methodName, methodID) => {
    const isPromise = promiseMethods && arrayContains(promiseMethods, methodID);
    const isSync = syncMethods && arrayContains(syncMethods, methodID);
    invariant(!isPromise || !isSync, 'Cannot have a method that is both async and a sync hook');
    const methodType = isPromise ? 'promise' : isSync ? 'sync' : 'async';
    module[methodName] = genMethod(moduleID, methodID, methodType);
  });
  Object.assign(module, constants);

  if (__DEV__) {
    BatchedBridge.createDebugLookup(moduleID, moduleName, methods);
  }

  return { name: moduleName, module };

若是存在方法值,则会遍历方法,调用 genMethod 来生成方法对应的信息:

之后便是将其赋值给 NativeModule:

const info = genModule(config, moduleID);
if (!info) {
  return;
}

if (info.module) {
  NativeModules[info.name] = info.module;
}

3.__fbBatchedBridgeConfig 的生成及赋值

在上述2的步骤中,其中 config 数组信息对应着 bridgeConfig 的属性 remoteModuleConfig。而 bridgeConfig 则是 global.__fbBatchedBridgeConfig 的值。

这个值则是在 JS 引擎的 ProxyExecutor 执行 loadApplicationScript 中进行赋值的:

void ProxyExecutor::loadApplicationScript(
    std::unique_ptr<const JSBigString>,
    std::string sourceURL) {

  folly::dynamic nativeModuleConfig = folly::dynamic::array;

  {
    SystraceSection s("collectNativeModuleDescriptions");
    auto moduleRegistry = m_delegate->getModuleRegistry();
    for (const auto& name : moduleRegistry->moduleNames()) {
      auto config = moduleRegistry->getConfig(name);
      nativeModuleConfig.push_back(config ? config->config : nullptr);
    }
  }

  folly::dynamic config =
    folly::dynamic::object
    ("remoteModuleConfig", std::move(nativeModuleConfig));

  {
    SystraceSection t("setGlobalVariable");
    setGlobalVariable(
      "__fbBatchedBridgeConfig",
      folly::make_unique<JSBigStdString>(folly::toJson(config)));
  }

  static auto loadApplicationScript =
    jni::findClassStatic(EXECUTOR_BASECLASS)->getMethod<void(jstring)>("loadApplicationScript");

  // The proxy ignores the script data passed in.

  loadApplicationScript(
    m_executor.get(),
    jni::make_jstring(sourceURL).get());
  // We can get pending calls here to native but the queue will be drained when
  // we launch the application.
}

这里利用 ExecutorDelegate 得到 ModuleRegistry,通过其获取到 moduleNames,进行遍历获取到每个 module 对应的 config 信息,将这些 config 信息放置在 nativeModuleConfig 数组中,并将其封装为 remoteModuleConfig 对象,调用 setGlobalVariable 方法,将此对象存放至名为 __fbBatchedBridgeConfig 的变量中。

4.JNI NativeModule 的注册

在上述过程中,就有个问题,NativeModule 是如何存在于 ModuleRegistry 中的?

首先暴露 JNI 相关的方法是在 OnLoad.cpp 的 JNI_OnLoad 方法中:

    extern "C" JNIEXPORT jint JNI_OnLoad(JavaVM *vm, void *reserved) {
      return initialize(vm, [] {
          gloginit::initialize();
          JSCJavaScriptExecutorHolder::registerNatives();
          ProxyJavaScriptExecutorHolder::registerNatives();
          CatalystInstanceImpl::registerNatives();
          CxxModuleWrapperBase::registerNatives();
          CxxModuleWrapper::registerNatives();
          JCxxCallbackImpl::registerNatives();
          NativeArray::registerNatives();
          ReadableNativeArray::registerNatives();
          WritableNativeArray::registerNatives();
          NativeMap::registerNatives();
          ReadableNativeMap::registerNatives();
          WritableNativeMap::registerNatives();
          ReadableNativeMapKeySetIterator::registerNatives();

#ifdef WITH_INSPECTOR
          JInspector::registerNatives();
#endif
      });

其调用了 CatalystInstanceImpl::registerNatives() 方法:

void CatalystInstanceImpl::registerNatives() {
  registerHybrid({
                   makeNativeMethod("initHybrid", CatalystInstanceImpl::initHybrid),
                   makeNativeMethod("initializeBridge",
                                    CatalystInstanceImpl::initializeBridge),
                   makeNativeMethod("jniExtendNativeModules",
                                    CatalystInstanceImpl::extendNativeModules),
                   makeNativeMethod("jniSetSourceURL",
                                    CatalystInstanceImpl::jniSetSourceURL),
                   makeNativeMethod("jniRegisterSegment",
                                    CatalystInstanceImpl::jniRegisterSegment),
                   makeNativeMethod("jniLoadScriptFromAssets",
                                    CatalystInstanceImpl::jniLoadScriptFromAssets),
                   makeNativeMethod("jniLoadScriptFromFile",
                                    CatalystInstanceImpl::jniLoadScriptFromFile),
                   makeNativeMethod("jniCallJSFunction",
                                    CatalystInstanceImpl::jniCallJSFunction),
                   makeNativeMethod("jniCallJSCallback",
                                    CatalystInstanceImpl::jniCallJSCallback),
                   makeNativeMethod("setGlobalVaeeriable",
                                    CatalystInstanceImpl::setGlobalVariable),
                   makeNativeMethod("getJavaScriptContext",
                                    CatalystInstanceImpl::getJavaScriptContext),
                   makeNativeMethod("jniHandleMemoryPressure",
                                    CatalystInstanceImpl::handleMemoryPressure),
                 });

  JNativeRunnable::registerNatives();
}

可以看到 CatalystInstanceImpl::initializeBridge 方法绑定为 initializeBridge JNI 方法,而此 JNI 方法的调用则是在初始化 CatalystInstanceImpl 的构造函数中:

  private CatalystInstanceImpl(
      final ReactQueueConfigurationSpec reactQueueConfigurationSpec,
      final JavaScriptExecutor jsExecutor,
      final NativeModuleRegistry nativeModuleRegistry,
      final JSBundleLoader jsBundleLoader,
      NativeModuleCallExceptionHandler nativeModuleCallExceptionHandler) {
    Log.d(ReactConstants.TAG, "Initializing React Xplat Bridge.");
    mHybridData = initHybrid();

    mReactQueueConfiguration = ReactQueueConfigurationImpl.create(
        reactQueueConfigurationSpec,
        new NativeExceptionHandler());
    mBridgeIdleListeners = new CopyOnWriteArrayList<>();
    mNativeModuleRegistry = nativeModuleRegistry;
    mJSModuleRegistry = new JavaScriptModuleRegistry();
    mJSBundleLoader = jsBundleLoader;
    mNativeModuleCallExceptionHandler = nativeModuleCallExceptionHandler;
    mNativeModulesQueueThread = mReactQueueConfiguration.getNativeModulesQueueThread();
    mTraceListener = new JSProfilerTraceListener(this);

    Log.d(ReactConstants.TAG, "Initializing React Xplat Bridge before initializeBridge");
    initializeBridge(
      new BridgeCallback(this),
      jsExecutor,
      mReactQueueConfiguration.getJSQueueThread(),
      mNativeModulesQueueThread,
      mNativeModuleRegistry.getJavaModules(this),
      mNativeModuleRegistry.getCxxModules());
    Log.d(ReactConstants.TAG, "Initializing React Xplat Bridge after initializeBridge");

    mJavaScriptContextHolder = new JavaScriptContextHolder(getJavaScriptContext());
  }

其中调用的 initializeBridge 调用的 Native 方法如下:

void CatalystInstanceImpl::initializeBridge(
  jni::alias_ref <ReactCallback::javaobject> callback,
  // This executor is actually a factory holder.
  JavaScriptExecutorHolder *jseh,
  jni::alias_ref <JavaMessageQueueThread::javaobject> jsQueue,
  jni::alias_ref <JavaMessageQueueThread::javaobject> nativeModulesQueue,
  jni::alias_ref <jni::JCollection<JavaModuleWrapper::javaobject>::javaobject> javaModules,
  jni::alias_ref <jni::JCollection<ModuleHolder::javaobject>::javaobject> cxxModules) {

  moduleMessageQueue_ = std::make_shared<JMessageQueueThread>(nativeModulesQueue);


  moduleRegistry_ = std::make_shared<ModuleRegistry>(
    buildNativeModuleList(
      std::weak_ptr<Instance>(instance_),
      javaModules,
      cxxModules,
      moduleMessageQueue_));

  instance_->initializeBridge(
    folly::make_unique<JInstanceCallback>(
      callback,
      moduleMessageQueue_),
    jseh->getExecutorFactory(),
    folly::make_unique<JMessageQueueThread>(jsQueue),
    moduleRegistry_);
}

其中实现初始化了 moduleRegistry,这里调用了 buildNativeModuleList 方法,其对应这 ModuleRegistryBuilder 类:

std::vector <std::unique_ptr<NativeModule>> buildNativeModuleList(
  std::weak_ptr <Instance> winstance,
  jni::alias_ref <jni::JCollection<JavaModuleWrapper::javaobject>::javaobject> javaModules,
  jni::alias_ref <jni::JCollection<ModuleHolder::javaobject>::javaobject> cxxModules,
  std::shared_ptr <MessageQueueThread> moduleMessageQueue) {
  std::vector <std::unique_ptr<NativeModule>> modules;
  if (javaModules) {
    for (const auto &jm : *javaModules) {
      modules.emplace_back(folly::make_unique<JavaNativeModule>(
        winstance, jm, moduleMessageQueue));
    }
  }
  if (cxxModules) {
    for (const auto &cm : *cxxModules) {
      modules.emplace_back(folly::make_unique<CxxNativeModule>(
        winstance, cm->getName(), cm->getProvider(), moduleMessageQueue));
    }
  }
  return modules;
}

这里则会遍历 javaModules 和延迟加载的 cxxModules,全部存放于类型为 NativeModule 的数组中,由此得到 ModuleRegistry 中,这样便可生成 config 信息:

folly::Optional <ModuleConfig> ModuleRegistry::getConfig(const std::string &name) {
  SystraceSection s("ModuleRegistry::getConfig", "module", name);

  // Initialize modulesByName_
  if (modulesByName_.empty() && !modules_.empty()) {
    moduleNames();
  }

  auto it = modulesByName_.find(name);

  if (it == modulesByName_.end()) {
    if (unknownModules_.find(name) != unknownModules_.end()) {
      return nullptr;
    }
    if (!moduleNotFoundCallback_ ||
        !moduleNotFoundCallback_(name) ||
        (it = modulesByName_.find(name)) == modulesByName_.end()) {
      unknownModules_.insert(name);
      return nullptr;
    }
  }
  size_t index = it->second;

  CHECK(index < modules_.size());
  NativeModule *module = modules_[index].get();

  // string name, object constants, array methodNames (methodId is index), [array promiseMethodIds], [array syncMethodIds]
  folly::dynamic config = folly::dynamic::array(name);

  {
    SystraceSection s_("getConstants");
    config.push_back(module->getConstants());
  }

  {
    SystraceSection s_("getMethods");
    std::vector <MethodDescriptor> methods = module->getMethods();

    folly::dynamic methodNames = folly::dynamic::array;
    folly::dynamic promiseMethodIds = folly::dynamic::array;
    folly::dynamic syncMethodIds = folly::dynamic::array;

    for (auto &descriptor : methods) {
      // TODO: #10487027 compare tags instead of doing string comparison?
      methodNames.push_back(std::move(descriptor.name));
      if (descriptor.type == "promise") {
        promiseMethodIds.push_back(methodNames.size() - 1);
      } else if (descriptor.type == "sync") {
        syncMethodIds.push_back(methodNames.size() - 1);
      }
    }

    if (!methodNames.empty()) {
      config.push_back(std::move(methodNames));
      if (!promiseMethodIds.empty() || !syncMethodIds.empty()) {
        config.push_back(std::move(promiseMethodIds));
        if (!syncMethodIds.empty()) {
          config.push_back(std::move(syncMethodIds));
        }
      }
    }
  }

  if (config.size() == 2 && config[1].empty()) {
    // no constants or methods
    return nullptr;
  } else {
    return ModuleConfig{index, config};
  }
}

其根据一个 name 值,从 modules 中获取到指定的 module,进行读取里面的值,得到 ModuleConfig 信息。

5.调用方法的暴露

这里提供了两个方法,来提供调用 Native 方法:

void ModuleRegistry::callNativeMethod(unsigned int moduleId, unsigned int methodId,
                                      folly::dynamic &¶ms, int callId) {
  if (moduleId >= modules_.size()) {
    throw std::runtime_error(
      folly::to<std::string>("moduleId ", moduleId, " out of range [0..", modules_.size(),
                             ")"));
  }
  modules_[moduleId]->invoke(methodId, std::move(params), callId);
}

MethodCallResult
ModuleRegistry::callSerializableNativeHook(unsigned int moduleId, unsigned int methodId,
                                           folly::dynamic &¶ms) {
  if (moduleId >= modules_.size()) {
    throw std::runtime_error(
      folly::to<std::string>("moduleId ", moduleId, "out of range [0..", modules_.size(),
                             ")"));
  }
  return modules_[moduleId]->callSerializableNativeHook(methodId, std::move(params));
}

这两个方法执行的是 module 对应的方法,对应类为 CxxNativeModule:

void
CxxNativeModule::invoke(unsigned int reactMethodId, folly::dynamic &¶ms, int callId) {
  if (reactMethodId >= methods_.size()) {
    throw std::invalid_argument(folly::to<std::string>("methodId ", reactMethodId,
                                                       " out of range [0..",
                                                       methods_.size(), "]"));
  }
  if (!params.isArray()) {
    throw std::invalid_argument(
      folly::to<std::string>("method parameters should be array, but are ",
                             params.typeName()));
  }

  CxxModule::Callback first;
  CxxModule::Callback second;

  const auto &method = methods_[reactMethodId];

  if (!method.func) {
    throw std::runtime_error(folly::to<std::string>("Method ", method.name,
                                                    " is synchronous but invoked asynchronously"));
  }

  if (params.size() < method.callbacks) {
    throw std::invalid_argument(folly::to<std::string>("Expected ", method.callbacks,
                                                       " callbacks, but only ",
                                                       params.size(),
                                                       " parameters provided"));
  }

  if (method.callbacks == 1) {
    first = convertCallback(makeCallback(instance_, params[params.size() - 1]));
  } else if (method.callbacks == 2) {
    first = convertCallback(makeCallback(instance_, params[params.size() - 2]));
    second = convertCallback(makeCallback(instance_, params[params.size() - 1]));
  }

  params.resize(params.size() - method.callbacks);


  messageQueueThread_->runOnQueue(
    [method, params = std::move(params), first, second, callId]() {
#ifdef WITH_FBSYSTRACE
        if (callId != -1) {
          fbsystrace_end_async_flow(TRACE_TAG_REACT_APPS, "native", callId);
        }
#endif
        SystraceSection s(method.name.c_str());
        try {
          method.func(std::move(params), first, second);
        } catch (const facebook::xplat::JsArgumentException &ex) {
          throw;
        } catch (std::exception &e) {
          LOG(ERROR) << "std::exception. Method call " << method.name.c_str() << " failed: "
                     << e.what();
          std::terminate();
        } catch (std::string &error) {
          LOG(ERROR) << "std::string. Method call " << method.name.c_str() << " failed: "
                     << error.c_str();
          std::terminate();
        } catch (...) {
          LOG(ERROR) << "Method call " << method.name.c_str() << " failed. unknown error";
          std::terminate();
        }
    });
}

在 invoke 方法中,会解析参数,确定 callback 的数量,之后便是在指定的 MessageQueue 中执行方法的调用。同理:

MethodCallResult
CxxNativeModule::callSerializableNativeHook(unsigned int hookId, folly::dynamic &&args) {
  if (hookId >= methods_.size()) {
    throw std::invalid_argument(
      folly::to<std::string>("methodId ", hookId, " out of range [0..", methods_.size(),
                             "]"));
  }

  const auto &method = methods_[hookId];

  if (!method.syncFunc) {
    throw std::runtime_error(
      folly::to<std::string>("Method ", method.name,
                             " is asynchronous but invoked synchronously"));
  }

  return method.syncFunc(std::move(args));
}

callSerialiableNativeHook 则是进行方法同步的调用。

Native 调用 JS

将 JS 端定义的对象方法来提供给 Native 端调用。

1.步骤

  • 1)Native端声明定义一个实现 JavaScriptModule 接口的类
  • 2) 在 JS 端,实现定义此 Module,并将此对象添加注册至 MessageQueue 中.

在 MessageQueue.js 中,提供了两个用来注册 Module 的方法:

registerCallableModule(name: string, module: Object)

以及懒加载方法:

registerLazyCallableModule(name: string, factory: void => Object)

2.实现

1.定义 JavaScriptModule

在 Native 中定义了 JavaScriptModule ,代表了在 JS 端定义的 JS 方法。

若需要调用 JS 端的方法时,使用 CatalystInstance 接口的 getJSModule 方法:

<T extends JavaScriptModule> T getJSModule(Class<T> jsInterface);

如调用 AppRegistryrunApplication 方法:

catalystInstance.getJSModule(AppRegistry.class).runApplication(jsAppModuleName, appParams);

2.调用过程

获取 JavaScriptModule 的方法,则是在 JavaScriptModuleRegistry 了中,其对 JavaScriptModule 类的获取,做了缓存策略,这里使用了代理调用,来改变其方法的调用。具体便是通过 JavaScriptModuleInvocationHandler 的 invoke 方法,来通过 CatalystInstance 来完成真正 JS 方法的调用:

@Override
public @Nullable Object invoke(Object proxy, Method method, @Nullable Object[] args) throws Throwable {
NativeArray jsArgs = args != null
  ? Arguments.fromJavaArgs(args)
  : new WritableNativeArray();
  mCatalystInstance.callFunction(getJSModuleName(), method.getName(), jsArgs);
  return null;
}

关于 CatalystInstance 在 C 层的实现,对应着 CatalystInstanceImpl.cpp 的类:

void CatalystInstanceImpl::jniCallJSFunction(std::string module, std::string method, NativeArray* arguments) {
  instance_->callJSFunction(std::move(module),
                            std::move(method),
                            arguments->consume());
}

instance_ 是指 Instance.cpp 类的实例,会调用到:

void Instance::callJSFunction(std::string &&module, std::string &&method,
                              folly::dynamic &¶ms) {
  callback_->incrementPendingJSCalls();
  nativeToJsBridge_->callFunction(std::move(module), std::move(method),
                                  std::move(params));
}

其调用的便是 NativeToJsBridge.cpp 的 callFunction 方法:

void NativeToJsBridge::callFunction(
    std::string&& module,
    std::string&& method,
    folly::dynamic&& arguments) {
  int systraceCookie = -1;

  ...

  runOnExecutorQueue([module = std::move(module), method = std::move(method), arguments = std::move(arguments), systraceCookie]
    (JSExecutor* executor) {
      ...
       // This is safe because we are running on the executor's thread: it won't
      // destruct until after it's been unregistered (which we check above) and
      // that will happen on this thread
      executor->callFunction(module, method, arguments);
    });
}

而 executor 便指的是 JSCExecutor.cpp 类:

 void JSCExecutor::callFunction(const std::string& moduleId, const std::string& methodId, const folly::dynamic& arguments) {
   SystraceSection s("JSCExecutor::callFunction");
   // This weird pattern is because Value is not default constructible.
   // The lambda is inlined, so there's no overhead.
   auto result = [&] {
     JSContextLock lock(m_context);
     try {
       if (!m_callFunctionReturnResultAndFlushedQueueJS) {
         bindBridge();
       }
       return m_callFunctionReturnFlushedQueueJS->callAsFunction({
         Value(m_context, String::createExpectingAscii(m_context, moduleId)),
         Value(m_context, String::createExpectingAscii(m_context, methodId)),
         Value::fromDynamic(m_context, std::move(arguments))
       });
     } catch (...) {
       std::throw_with_nested(
                              std::runtime_error("Error calling " + moduleId + "." + methodId));
     }
   }();
   callNativeModules(std::move(result));
 }

其中的 m_callFunctionReturnResultAndFlushedQueueJS 是在 bindBridge 过程中,从 JS 全局对象 __fbBatchedBridge 中进行获取读到的:

 void JSCExecutor::bindBridge() throw(JSException) {
   SystraceSection s("JSCExecutor::bindBridge");
   std::call_once(m_bindFlag, [this] {
     auto global = Object::getGlobalObject(m_context);
     auto batchedBridgeValue = global.getProperty("__fbBatchedBridge");
     if (batchedBridgeValue.isUndefined()) {
       auto requireBatchedBridge = global.getProperty("__fbRequireBatchedBridge");
       if (!requireBatchedBridge.isUndefined()) {
         batchedBridgeValue = requireBatchedBridge.asObject().callAsFunction({});
       }
       if (batchedBridgeValue.isUndefined()) {
         throw JSException("Could not get BatchedBridge, make sure your bundle is packaged correctly");
       }
     }

     auto batchedBridge = batchedBridgeValue.asObject();
     m_callFunctionReturnFlushedQueueJS = batchedBridge.getProperty("callFunctionReturnFlushedQueue").asObject();
     m_invokeCallbackAndReturnFlushedQueueJS = batchedBridge.getProperty("invokeCallbackAndReturnFlushedQueue").asObject();
     m_flushedQueueJS = batchedBridge.getProperty("flushedQueue").asObject();
     m_callFunctionReturnResultAndFlushedQueueJS = batchedBridge.getProperty("callFunctionReturnResultAndFlushedQueue").asObject();
   });
 }

而此 __fbBatchedBridge 的定义则是在 JS 端的 BatchedBridge.js 文件中:

'use strict';

const MessageQueue = require('MessageQueue');

const BatchedBridge = new MessageQueue();

// Wire up the batched bridge on the global object so that we can call into it.
// Ideally, this would be the inverse relationship. I.e. the native environment
// provides this global directly with its script embedded. Then this module
// would export it. A possible fix would be to trim the dependencies in
// MessageQueue to its minimal features and embed that in the native runtime.

Object.defineProperty(global, '__fbBatchedBridge', {
  configurable: true,
  value: BatchedBridge,
});

module.exports = BatchedBridge;

这里其值则是对应着 MessageQueue,其管理着 JS 端的 function 的注册调用逻辑,其中 callFunctionReturnResultAndFlushedQueue

  callFunctionReturnFlushedQueue(module: string, method: string, args: any[]) {
    this.__guard(() => {
      this.__callFunction(module, method, args);
    });

    return this.flushedQueue();
  }

__callFunction 方法的调用:

__callFunction(module: string, method: string, args: any[]): any {
 this._lastFlush = new Date().getTime();
 this._eventLoopStartTime = this._lastFlush;
 Systrace.beginEvent(`${module}.${method}()`);
 if (this.__spy) {
   this.__spy({type: TO_JS, module, method, args});
 }
 const moduleMethods = this.getCallableModule(module);
 invariant(
   !!moduleMethods,
   'Module %s is not a registered callable module (calling %s)',
   module,
   method,
 );
 invariant(
   !!moduleMethods[method],
   'Method %s does not exist on module %s',
   method,
   module,
 );
 const result = moduleMethods[method].apply(moduleMethods, args);
 Systrace.endEvent();
 return result;
}

其中的 getCallableModule 则是从 _lazyCallableModules 中进行获取,其注册则是通过 registerCallableModule 方法。以 AppRegistry.js 为例,在其文件的末尾则是调用了 BatchedBridge 进行注册:

BatchedBridge.registerCallableModule('AppRegistry', AppRegistry);
原文  http://alighters.com/blog/2018/07/09/react-native-message-theory/
正文到此结束
Loading...