摘要：流程圖大概如下的源碼比較簡單，而在執行更新的過程比較復雜。這段代碼的核心就是調用，然后對老的屬性和狀態存一下，新的更新一下而已。最后是把執行推入的隊列中，等待組件的更新。最后調用統一修改的屬性。

原文鏈接地址：https://github.com/Nealyang%EF%BC%9A%E7%8A%B6%E6%80%81%E3%80%81%E5%B1%9E%E6%80%A7%E6%9B%B4%E6%96%B0%20-%3E%20setState.md) 轉載請注明出處

此次分析setState基于0.3版本，實現比較簡單，后續會再分析目前使用的版本以及事務機制。

流程圖大概如下

setState的源碼比較簡單，而在執行更新的過程比較復雜。我們直接跟著源碼一點一點屢清楚。

ReactCompositeComponent.js

  /**
   * Sets a subset of the state. Always use this or `replaceState` to mutate
   * state. You should treat `this.state` as immutable.
   *
   * There is no guarantee that `this.state` will be immediately updated, so
   * accessing `this.state` after calling this method may return the old value.
   *
   * @param {object} partialState Next partial state to be merged with state.
   * @final
   * @protected
   */
  setState: function(partialState) {
    // Merge with `_pendingState` if it exists, otherwise with existing state.
    this.replaceState(merge(this._pendingState || this.state, partialState));
  },

注釋部分說的很明確，setState后我們不能夠立即拿到我們設置的值。

而這段代碼也非常簡單，就是將我們傳入的state和this._pendingState做一次merge，merge的代碼在util.js下

var merge = function(one, two) {
  var result = {};
  mergeInto(result, one);
  mergeInto(result, two);
  return result;
};

function mergeInto(one, two) {
  checkMergeObjectArg(one);
  if (two != null) {
    checkMergeObjectArg(two);
    for (var key in two) {
      if (!two.hasOwnProperty(key)) {
        continue;
      }
      one[key] = two[key];
    }
  }
}

  checkMergeObjectArgs: function(one, two) {
    mergeHelpers.checkMergeObjectArg(one);
    mergeHelpers.checkMergeObjectArg(two);
  },

  /**
   * @param {*} arg
   */
  checkMergeObjectArg: function(arg) {
    throwIf(isTerminal(arg) || Array.isArray(arg), ERRORS.MERGE_CORE_FAILURE);
  },
  
  var isTerminal = function(o) {
  return typeof o !== "object" || o === null;
};

var throwIf = function(condition, err) {
  if (condition) {
    throw new Error(err);
  }
};

診斷代碼的邏輯非常簡單，其實功能就是Object.assign() ,但是從上面代碼我們可以看出react源碼中的function大多都具有小而巧的特點。

最終，將merge后的結果傳遞給replaceState

replaceState: function(completeState) {
    var compositeLifeCycleState = this._compositeLifeCycleState;
    invariant(
      this._lifeCycleState === ReactComponent.LifeCycle.MOUNTED ||
      compositeLifeCycleState === CompositeLifeCycle.MOUNTING,
      "replaceState(...): Can only update a mounted (or mounting) component."
    );
    invariant(
      compositeLifeCycleState !== CompositeLifeCycle.RECEIVING_STATE &&
      compositeLifeCycleState !== CompositeLifeCycle.UNMOUNTING,
      "replaceState(...): Cannot update while unmounting component or during " +
      "an existing state transition (such as within `render`)."
    );

    this._pendingState = completeState;

    // Do not trigger a state transition if we are in the middle of mounting or
    // receiving props because both of those will already be doing this.
    if (compositeLifeCycleState !== CompositeLifeCycle.MOUNTING &&
        compositeLifeCycleState !== CompositeLifeCycle.RECEIVING_PROPS) {
      this._compositeLifeCycleState = CompositeLifeCycle.RECEIVING_STATE;

      var nextState = this._pendingState;
      this._pendingState = null;

      var transaction = ReactComponent.ReactReconcileTransaction.getPooled();
      transaction.perform(
        this._receivePropsAndState,
        this,
        this.props,
        nextState,
        transaction
      );
      ReactComponent.ReactReconcileTransaction.release(transaction);

      this._compositeLifeCycleState = null;
    }
  },

撇開50% 判斷warning代碼不說，從上面代碼我們可以看出，只有在componsiteLifeState不等于mounting和receiving_props 時，才會調用 _receivePropsAndState函數來更新組件。

我們可以演示下：

var ExampleApplication = React.createClass({
      getInitialState() {
        return {}
      },
      componentWillMount() {
        this.setState({
          a: 1,
        })
        console.log("componentWillMount", this.state.a)
        this.setState({
          a: 2,
        })
        console.log("componentWillMount", this.state.a)
        this.setState({
          a: 3,
        })
        console.log("componentWillMount", this.state.a)
        setTimeout(() => console.log("a5"), 0)
        setTimeout(() => console.log(this.state.a,"componentWillMount"))

        Promise.resolve("a4").then(console.log)
      },

      componentDidMount() {
        this.setState({
          a: 4,
        })
        console.log("componentDidMount", this.state.a)
        this.setState({
          a: 5,
        })
        console.log("componentDidMount", this.state.a)
        this.setState({
          a: 6,
        })
        console.log("componentDidMount", this.state.a)
      },
      render: function () {
        var elapsed = Math.round(this.props.elapsed / 100);
        var seconds = elapsed / 10 + (elapsed % 10 ? "" : ".0");
        var message =
          "React has been successfully running for " + seconds + " seconds.";
        return React.DOM.p(null, message);
      }
    });

所以以上結果我們可以看出，在componentWillMount生命周期內setState后this.state不會改變，在componentDidMount是正常的。因為在上一篇文章中我們也有說到，在mountComponent過程中，會把compositeLifeCycleState設置為MOUNTING狀態，在這個過程中，是不會執行receivePropsAndState的，所以this.state也就不會更新，同理，在receivePropsAndState的過程中，會把compositeLifeCycleState置成RECEIVING_PROPS狀態，也不會執行state更新以及render執行，在updateComponent過程中又執行了mountComponent函數，mountComponent函數調用了render函數。

而在現在我們使用16或者15版本中，我們發現：

componentDidMount() {
    this.setState({val: this.state.val + 1});
    console.log(this.state.val);    // 第 1 次 log

    this.setState({val: this.state.val + 1});
    console.log(this.state.val);    // 第 2 次 log

    setTimeout(() => {
      this.setState({val: this.state.val + 1});
      console.log(this.state.val);  // 第 3 次 log

      this.setState({val: this.state.val + 1});
      console.log(this.state.val);  // 第 4 次 log
    }, 0);
  }

最后打印的結果為：0，0，2，3

為什么有這樣呢？其實源于源碼中的這段代碼：

function enqueueUpdate(component) {
  ensureInjected();

  // Various parts of our code (such as ReactCompositeComponent"s
  // _renderValidatedComponent) assume that calls to render aren"t nested;
  // verify that that"s the case. (This is called by each top-level update
  // function, like setProps, setState, forceUpdate, etc.; creation and
  // destruction of top-level components is guarded in ReactMount.)

  if (!batchingStrategy.isBatchingUpdates) {
    batchingStrategy.batchedUpdates(enqueueUpdate, component);
    return;
  }

  dirtyComponents.push(component);
}

因為這里涉及到事務的概念、批量更新以及benchUpdate等，在我們目前分析的版本中還未迭代上去，后面我們會跟著版本升級慢慢說道。

首先我們知道，屬性的更新必然是由于state的更新，所以其實組件屬性的更新流程就是setState執行更新的延續，換句話說，也就是setState才能出發組件屬性的更新，源碼里就是我在處理state更新的時候，順帶檢測了屬性的更新。所以這段源碼的開始，還是從setState中看

  _receivePropsAndState: function(nextProps, nextState, transaction) {
    if (!this.shouldComponentUpdate ||
        this.shouldComponentUpdate(nextProps, nextState)) {
      this._performComponentUpdate(nextProps, nextState, transaction);
    } else {
      this.props = nextProps;
      this.state = nextState;
    }
  },

代碼非常的簡單，一句話解釋：當shouldComponentUpdate為true時，則執行更新操作。

  _performComponentUpdate: function(nextProps, nextState, transaction) {
    var prevProps = this.props;
    var prevState = this.state;

    if (this.componentWillUpdate) {
      this.componentWillUpdate(nextProps, nextState, transaction);
    }

    this.props = nextProps;
    this.state = nextState;

    this.updateComponent(transaction);

    if (this.componentDidUpdate) {
      transaction.getReactOnDOMReady().enqueue(
        this,
        this.componentDidUpdate.bind(this, prevProps, prevState)
      );
    }
  },

這段代碼的核心就是調用this.updateComponent，然后對老的屬性和狀態存一下，新的更新一下而已。如果存在componentWillUpdate就執行一下，然后走更新流程。最后是把執行componentDidUpdate推入getReactOnDOMReady的隊列中，等待組件的更新。

  _renderValidatedComponent: function() {
    ReactCurrentOwner.current = this;
    var renderedComponent = this.render();
    ReactCurrentOwner.current = null;
    return renderedComponent;
  },
  ...
  ...
  updateComponent: function(transaction) {
    var currentComponent = this._renderedComponent;
    var nextComponent = this._renderValidatedComponent();
    if (currentComponent.constructor === nextComponent.constructor) {
      if (!nextComponent.props.isStatic) {
        currentComponent.receiveProps(nextComponent.props, transaction);
      }
    } else {
      var thisID = this._rootNodeID;
      var currentComponentID = currentComponent._rootNodeID;
      currentComponent.unmountComponent();
      var nextMarkup = nextComponent.mountComponent(thisID, transaction);
      ReactComponent.DOMIDOperations.dangerouslyReplaceNodeWithMarkupByID(
        currentComponentID,
        nextMarkup
      );
      this._renderedComponent = nextComponent;
    }
  },

這里我們直接看updateComponent更新流程，首先獲取當前render函數的組件，然后獲取下一次render函數的組件，_renderValidatedComponent就是獲取下一次的render組件。 通過Constructor來判斷組件是否相同，如果相同且組件為非靜態，則更新組件的屬性，否則卸載當前組件，然后重新mount下一個render組件并且直接暴力更新。

接著會調用render組件的receiveProps方法，其實一開始這個地方我也是非常困惑的，this指向傻傻分不清楚，后來經過各種查閱資料知道，它其實是一個多態方法，如果是復合組件，則執行ReactCompositeComponent.receiveProps，如果是原生組件，則執行ReactNativeComponent.receiveProps。源碼分別如下：

  receiveProps: function(nextProps, transaction) {
    if (this.constructor.propDeclarations) {
      this._assertValidProps(nextProps);
    }
    ReactComponent.Mixin.receiveProps.call(this, nextProps, transaction);

    this._compositeLifeCycleState = CompositeLifeCycle.RECEIVING_PROPS;
    if (this.componentWillReceiveProps) {
      this.componentWillReceiveProps(nextProps, transaction);
    }
    this._compositeLifeCycleState = CompositeLifeCycle.RECEIVING_STATE;
    var nextState = this._pendingState || this.state;
    this._pendingState = null;
    this._receivePropsAndState(nextProps, nextState, transaction);
    this._compositeLifeCycleState = null;
  },

有人可能注意到這里的this._receivePropsAndState函數，這不是剛才調用過么？怎么又調用一遍？沒錯，調用這個的this已經是currentComponent了，并不是上一個this。currentComponent是當前組件的render組件，也就是當前組件的子組件。子組件同樣也可能是復合組件或者原生組件。正式通過這種多態的方式，遞歸的解析每級嵌套組件。最終完成從當前組件到下面的所有葉子節點的樹更新。

其實話說回來，compositeComponent最終還是會遍歷遞歸到解析原生組件，通過我們整體瀏覽下ReactNativeComponent.js代碼可以看出。

我們先從 receiveProps方法開始看

  receiveProps: function(nextProps, transaction) {
    assertValidProps(nextProps);
    ReactComponent.Mixin.receiveProps.call(this, nextProps, transaction);
    this._updateDOMProperties(nextProps);
    this._updateDOMChildren(nextProps, transaction);
    this.props = nextProps;
  },
  
  function assertValidProps(props) {
  if (!props) {
    return;
  }
  var hasChildren = props.children != null ? 1 : 0;
  var hasContent = props.content != null ? 1 : 0;
  var hasInnerHTML = props.dangerouslySetInnerHTML != null ? 1 : 0;
}

刪除安全警告和注釋其實代碼非常簡答，首先assertValidProps就是校驗props是否合法的，更新屬性的方法就是_updateDOMProperties

_updateDOMProperties: function(nextProps) {
    var lastProps = this.props;
    for (var propKey in nextProps) {
      var nextProp = nextProps[propKey];
      var lastProp = lastProps[propKey];
      //判斷新老屬性中的值是否相等
      if (!nextProps.hasOwnProperty(propKey) || nextProp === lastProp) {
        continue;
      }
      //如果是style樣式，遍歷新style，如果去舊style不相同，則把變化的存入styleUpdates對象中。最后調用 updateStylesByID 統一修改dom的style屬性。
      if (propKey === STYLE) {
        if (nextProp) {
          nextProp = nextProps.style = merge(nextProp);
        }
        var styleUpdates;
        for (var styleName in nextProp) {
          if (!nextProp.hasOwnProperty(styleName)) {
            continue;
          }
          if (!lastProp || lastProp[styleName] !== nextProp[styleName]) {
            if (!styleUpdates) {
              styleUpdates = {};
            }
            styleUpdates[styleName] = nextProp[styleName];
          }
        }
        if (styleUpdates) {
          ReactComponent.DOMIDOperations.updateStylesByID(
            this._rootNodeID,
            styleUpdates
          );
        }
      } else if (propKey === DANGEROUSLY_SET_INNER_HTML) {
        var lastHtml = lastProp && lastProp.__html;
        var nextHtml = nextProp && nextProp.__html;
        if (lastHtml !== nextHtml) {
          ReactComponent.DOMIDOperations.updateInnerHTMLByID(//注意這里是innerHtml，所以dangerouslyInnerHTML會展示正常的HTML
            this._rootNodeID,
            nextProp
          );
        }
      } else if (propKey === CONTENT) {
        ReactComponent.DOMIDOperations.updateTextContentByID(//這里是innerText，所以content與children原封不動的把HTML代碼打印到頁面上
          this._rootNodeID,
          "" + nextProp
        );
      } else if (registrationNames[propKey]) {
        putListener(this._rootNodeID, propKey, nextProp);
      } else {
        ReactComponent.DOMIDOperations.updatePropertyByID(
          this._rootNodeID,
          propKey,
          nextProp
        );
      }
    }
  },

這里面方法沒有太多的hack技巧，非常的簡單直白，不多帶帶擰出來說，我直接寫到注釋里面了。

最后直接更新組件的屬性

  setValueForProperty: function(node, name, value) {
    if (DOMProperty.isStandardName[name]) {
      var mutationMethod = DOMProperty.getMutationMethod[name];
      if (mutationMethod) {
        mutationMethod(node, value);
      } else if (DOMProperty.mustUseAttribute[name]) {
        if (DOMProperty.hasBooleanValue[name] && !value) {
          node.removeAttribute(DOMProperty.getAttributeName[name]);
        } else {
          node.setAttribute(DOMProperty.getAttributeName[name], value);
        }
      } else {
        var propName = DOMProperty.getPropertyName[name];
        if (!DOMProperty.hasSideEffects[name] || node[propName] !== value) {
          node[propName] = value;
        }
      }
    } else if (DOMProperty.isCustomAttribute(name)) {
      node.setAttribute(name, value);
    }
  }

整體屬性更新的流程圖大概如下：

通篇讀完，是不是有種

react源碼中包含很多的點的知識，比如我們之前說的VDOM、包括后面要去學習dom-diff、事務、緩存等等，都是一個點，而但從一個點來切入難免有的會有些枯燥沒卵用，別急別急~