状态模式State

在组件构建过程中，某些对象的状态经常面临变化，如何对这些变化进行有效的管理？同时又维持高层模块的稳定？“状态变化”模式为这一问题提供了一种解决方案。

经典模式：State、Memento

动机（Motivation）

在软件构建过程中，某些对象的状态如何改变，其行为也随之而发生变化，比如文档处于只读状态，其支持的行为和读写状态支持的行为就可以完全不同。

如何在运行时根据对象的状态来透明地更改对象的行为？而不会为对象操作和状态转化之间引入耦合？

示例1

enum NetworkState
{Network_Open,Network_Close,Network_Connect,NetWork_Wait    // 新增状态
};class NetworkProcessor {NetworkState state_;public:void Operation1() {if (Network_Open == state_) {//***state_ = Network_Close;}else if(Network_Close == state_){//...state_ = Network_Connect;}else if (Network_Connect == state_) {//$$$state_ = Network_Open;}// 变化else if (NetWork_Wait == state_) {//---}}void Operation2() {if (Network_Open == state_) {//***state_ = Network_Connect;}else if (Network_Close == state_){//...state_ = Network_Open;}else if (Network_Connect == state_) {//$$$state_ = Network_Close;}// 变化else if (NetWork_Wait == state_) {//---}}void Operation3() {//...}
};

示例2

class NetworkState {
public:NetworkState* pNext;virtual void Operation1() = 0;virtual void Operation2() = 0;virtual void Operation3() = 0;virtual ~NetworkState() {}
};class OpenState : public NetworkState {static NetworkState* m_instance;public:static NetworkState* getInstance() {if (nullptr == m_instance) {m_instance = new OpenState();}return m_instance;}virtual void Operation1() {//***pNext = CloseState::getInstance();}virtual void Operation2() {//...pNext = ConnectState::getInstance();}virtual void Operation3() {//$$$pNext = OpenState::getInstance();}
};class CloseState : public NetworkState {static NetworkState* m_instance;public:static NetworkState* getInstance() {if (nullptr == m_instance) {m_instance = new CloseState();}return m_instance;}virtual void Operation1() {//***pNext = ConnectState::getInstance();}virtual void Operation2() {//...pNext = OpenState::getInstance();}virtual void Operation3() {//$$$pNext = CloseState::getInstance();}
};class ConnectState : public NetworkState {static NetworkState* m_instance;public:static NetworkState* getInstance() {if (nullptr == m_instance) {m_instance = new ConnectState();}return m_instance;}virtual void Operation1() {//***pNext = OpenState::getInstance();}virtual void Operation2() {//...pNext = CloseState::getInstance();}virtual void Operation3() {//$$$pNext = ConnectState::getInstance();}
};// 扩展
class WaitState : public NetworkState {//...
};class NetworkProcessor {NetworkState* state_;public:NetworkProcessor(NetworkState* state) {this->state_ = state;}void Operation1() {//...state_->Operation1();state_ = state_->pNext;//...}void Operation2() {//...state_->Operation2();state_ = state_->pNext;//...}void Operation3() {//...state_->Operation3();state_ = state_->pNext;//...}};

要点总结

State模式将所有与一个特定状态的相关的行为都放入一个State的子类对象中，在对象状态切换时，切换相应的对象；但同时维持State的接口，这样实现了具体操作与状态转换之间的解耦。

为不同的状态引入不同的对象使得状态转换变得更加明确，而且可怀疑保证不会出现状态不一致的情况，因为转换是原子性的---即要么彻底转换过来，要么不转换。

如果State对象没有实例变量，那么各个上下文可怀疑共享同一个State对象，从而节省对象开销。