封装stream,在读写stream时提供事件通知_.Net教程

编辑Tag赚U币
教程Tag:暂无Tag,欢迎添加,赚取U币!

推荐:实例开发:ASP.NET创建网络相册
在现在的数码时代,我们会经常拍摄一些相片以供留念,而随着数码照片的增多,往往需要很好地管理这些照片,以便更好地查阅留念。现在网上有不少的电子相册,都能很好的实现这些功能,那我们能否

前阵子的工作涉及一些网络编程,使用了面向流的方式做传输数据。在代码过程中,遇到一个新需求就是要统计流量。其实最简单的办法就时在读写流的地方增加代码,把功能增加上去就可以。但是我觉得那样对我原理的代码框架影响较大,基于尽量不影响原来的代码的考虑,我想到了Decorator设计模式。

先把代码贴出来,在做解释吧:

以下为引用的内容:
public class EventStream : Stream
{
public event EventHandler<FStreamDataEventArgs> OnBeforeRead;
public event EventHandler<FStreamDataEventArgs> OnBeforeWrite;

private Stream stream;
public EventStream(Stream stream)
{
if (stream == null) throw new ArgumentNullException("EventStream");
this.stream = stream;
}

[ ==== Stream members ==== ]#region [ ==== Stream members ==== ]
public override bool CanRead
{
get { return stream.CanRead; }
}

public override bool CanSeek
{
get { return stream.CanSeek; }
}

public override bool CanWrite
{
get { return stream.CanWrite; }
}

public override void Flush()
{
stream.Flush();
}

public override long Length
{
get { return stream.Length; }
}

public override long Position
{
get
{
return stream.Position;
}
set
{
stream.Position = value;
}
}

public override int Read(byte[] buffer, int offset, int count)
{
int readSize = stream.Read(buffer, offset, count);
if (OnBeforeRead != null)
OnBeforeRead(this, new FStreamDataEventArgs(buffer, offset, readSize));
return readSize;
}

public override long Seek(long offset, SeekOrigin origin)
{
return stream.Seek(offset, origin);
}

public override void SetLength(long value)
{
stream.SetLength(value);
}

public override void Write(byte[] buffer, int offset, int count)
{
if (OnBeforeWrite != null)
OnBeforeWrite(this, new FStreamDataEventArgs(buffer, offset, count));
stream.Write(buffer, offset, count);
}

public override IAsyncResult BeginRead(byte[] buffer, int offset, int count,
AsyncCallback callback, object state)
{
InternalAsyncState myState = new InternalAsyncState(
new FStreamDataEventArgs(buffer, offset, count), state);
AsyncCallback myCallback = new AsyncCallback(
new InternalCallback(OnBeforeRead, callback).Callback);
return new EventStreamAsyncResult(
stream.BeginRead(buffer, offset, count, myCallback, myState));
}

public override int EndRead(IAsyncResult asyncResult)
{
EventStreamAsyncResult esar = asyncResult as EventStreamAsyncResult;
if (esar != null)
return stream.EndRead(esar.InternalAsyncResult);
else
return stream.EndRead(asyncResult);
}

public override IAsyncResult BeginWrite(byte[] buffer, int offset, int count, AsyncCallback callback, object state)
{
InternalAsyncState myState = new InternalAsyncState(
new FStreamDataEventArgs(buffer, offset, count), state);
AsyncCallback myCallback = new AsyncCallback(
new InternalCallback(OnBeforeWrite, callback).Callback);
return new EventStreamAsyncResult(
stream.BeginWrite(buffer, offset, count, myCallback, myState));
}

public override void EndWrite(IAsyncResult asyncResult)
{
stream.EndWrite(asyncResult);
}

#endregion

private class InternalCallback
{
private AsyncCallback callback;
private EventHandler<FStreamDataEventArgs> internalHandler;

public InternalCallback(EventHandler<FStreamDataEventArgs> internalHandler, AsyncCallback callback)
{
this.internalHandler = internalHandler;
this.callback = callback;
}

internal void Callback(IAsyncResult asyncResult)
{
InternalAsyncState myState = asyncResult.AsyncState as InternalAsyncState;
if (internalHandler != null && myState != null)
internalHandler(this, myState.StreamDataEventArgs);
callback(new EventStreamAsyncResult(asyncResult));
}
}

private class InternalAsyncState
{
object state;
FStreamDataEventArgs streamDataEventArgs;

public object State
{
get { return state; }
}

public FStreamDataEventArgs StreamDataEventArgs
{
get { return streamDataEventArgs; }
}

public InternalAsyncState(FStreamDataEventArgs streamDataEventArgs, object state)
{
this.streamDataEventArgs = streamDataEventArgs;
this.state = state;
}
}

private class EventStreamAsyncResult : IAsyncResult
{
IAsyncResult ar;

public EventStreamAsyncResult(IAsyncResult ar)
{
if (ar == null) throw new ArgumentNullException("EventStreamAsyncResult");
this.ar = ar;
}
IAsyncResult Members#region IAsyncResult Members

public object AsyncState
{
get
{
InternalAsyncState myState = ar.AsyncState as InternalAsyncState;
if (myState != null)
return myState.State;
else
return ar.AsyncState;
}
}

internal IAsyncResult InternalAsyncResult
{
get { return ar; }
}

public System.Threading.WaitHandle AsyncWaitHandle
{
get { return ar.AsyncWaitHandle; }
}

public bool CompletedSynchronously
{
get { return ar.CompletedSynchronously; }
}

public bool IsCompleted
{
get { return ar.IsCompleted; }
}

#endregion
}
}

public class FStreamDataEventArgs : EventArgs
{
private byte[] buffer;
private int offset;
private int count;

public FStreamDataEventArgs(byte[] buffer, int offset, int count)
{
if(buffer == null) throw new ArgumentNullException("FStreamDataEventArgs");
if(offset count>buffer.Length) throw new ArgumentOutOfRangeException("FStreamDataEventArgs");

this.buffer = buffer;
this.offset = offset;
this.count = count;
}

/**//// <summary>
/// 数据缓存
/// </summary>
public byte[] Buffer
{
get { return buffer; }
}

/**//// <summary>
/// 数据开始位置
/// </summary>
public int Offset
{
get { return offset; }
}

/**//// <summary>
/// 数据长度
/// </summary>
public int Count
{
get { return count; }
}
}

刚开始以为很简单,事实上写下来还挺多行代码的,Decorator模式嘛,当然先继承stream,把stream本来该做的事情先完成了。这个很简单类里面包含一个内部的stream,stream该有的接口都由它来完成了。接下来就是增加两个事件,分别是OnBeforeRead、OnBeforeWrite。名字里面都有Before,其实我考虑到数据流都会通过这两个事件开放出来,你想做加密什么的都可以,当然也包括我想要的统计数据流量。

接下来就是在读写流的时候触发这两个事件就可以了。看看同步的Read、Write方法,简单的调用就可以了。
关键的地方就在于异步的读写。

我们先看看一般Stream的异步调用代码是怎么样的:

以下为引用的内容:

stream.BeginRead(buffer, 0, byte2read, new AsyncCallback(EndReadCallback), state);

private void EndReadCallback(IAsyncResult asyncResult)
{
object state = asyncResult.AsyncState;
nReadSize = stream.EndRead(asyncResult);
//
}

在不更改这个“client”代码的情况下,要怎么样在stream那边知道这里的确实读了多少数据呢?

显然在调用BeginRead的时候是不知道,那就只能对这个AsyncCallback做手脚了。可以预想到framework内部会在完成了Read的操作之后会调用AsyncCallback委托来通知结果。于是我就传一个我定义好的AsyncCallback委托给BeginRead。当然还要把“client”提供的AsyncCallback给包装起来,在做完我的事情(事件通知)之后,还是要把“client”要我办的事情给也给办了(调用"client"的AsyncCallback委托来通知结果)。

这就在实现了“在客户代码与framework之间插一脚”。

再来看看我是怎么做到事件通知的。首先要把我要的数据给传过去,于是有了InternalAsyncState,这里面要有我触发事件需要的事件参数,还应该要包括用户可能传入的state。具体大家看看InternalAsyncState的实现。

最后多考虑了一点就是,假如“client”代码不是像我写的那样,而是不断的通过检查Stream.BeginRead 方法返回的IAsyncResult的IsCompleted属性来确定是否Read完成的话,那我的代码就有问题了,我返回的IAsyncResult根本就不是原理的IAsyncResult了。EventStreamAsyncResult类就是为这个而写的。
下面是使用的代码:

以下为引用的内容:
public void GetResponseStream()
{
EventStream es = new EventStream(tcpClient.NetStream);
es.OnBeforeRead = new EventHandler<FStreamDataEventArgs>(EventStream_OnBeforeRead);
es.OnBeforeWrite = new EventHandler<FStreamDataEventArgs>(EventStream_OnBeforeWrite);
return es;
}

回头看看代码,其实都在用Decorator模式的思想,把原来的framework中的类都给包装起来,并在完成原来的功能之余另外加了自己的功能。

文笔一般,希望能对你有帮助。

分享:“黑盒测试管理”以外的编程经验片断
除了很好地进行“黑盒测试管理”,个人觉得下述更值得用功;毕竟“前除后略”! 1、架构和算法的可行性测试分析:主要包括性能、并发等方面 2、CodeReview:很简单,但

来源:模板无忧//所属分类:.Net教程/更新时间:2008-08-22
相关.Net教程