NetDataContractSerializer产生无效的XML

Question

我NetDataContractSerializer似乎混淆：在XML年底出现了两次：

<?xml version="1.0" encoding="utf-8"?> 
<Project xmlns:i="http://www.w3.org/2001/XMLSchema-instance" z:Id="1" 

[...] 
     <d2p1:anyType i:nil="true" /> 
     </d2p1:_items> 
     <d2p1:_size>2</d2p1:_size> 
     <d2p1:_version>2</d2p1:_version> 
    </d2p1:items> 
    </ProjectParts> 
    <ProjectPath z:Id="31">D:\t10\</ProjectPath> 
</Project>ze> 
       <d2p1:_version>3</d2p1:_version> 
      </d2p1:items> 
      <d2p1:_monitor xmlns:d7p1="http://schemas.datacontract.org/2004/07/System.Collections.ObjectModel" z:Id="33"> 
       <d7p1:_busyCount>0</d7p1:_busyCount> 
      </d2p1:_monitor> 
      </Elements> 
      <Project z:Ref="1" i:nil="true" xmlns="http://schemas.datacontract.org/2004/07/Modules.WorkspaceManager.Types" /> 
     </d2p1:anyType> 
     <d2p1:anyType i:nil="true" /> 
     <d2p1:anyType i:nil="true" /> 
     </d2p1:_items> 
     <d2p1:_size>2</d2p1:_size> 
     <d2p1:_version>2</d2p1:_version> 
    </d2p1:items> 
    </ProjectParts> 
    <ProjectPath z:Id="34">D:\t10\</ProjectPath> 
</Project> 

正如你可以看到，有一些严重的口吃怎么回事。它偶尔会发生，我不能重现错误。有任何想法吗？这是否可能是由正在写入的VS中打开的文件造成的？

序列化我的对象是这样的：

private void SerializeToFile(object objectToSerialize) 
    { 
     Stream stream = null; 

     try 
     { 
      stream = File.Open(_fileName, FileMode.OpenOrCreate, FileAccess.Write); 
      using (var writer = XmlWriter.Create(stream, new XmlWriterSettings { Indent = true })) 
      { 
       NetDataContractSerializer serializer = new NetDataContractSerializer(); 
       serializer.WriteObject(writer, objectToSerialize); 
      } 
     } 

     finally 
     { 
      if (stream != null) stream.Close(); 
     } 
    } 

而且序列化的类看起来像这样：

[DataContract(IsReference = true)] 
public class Project : IProject 
{ 
    [DataMember] public string ProjectPath { get; set; } 
    [DataMember] public string ProjectName { get; set; } 
    [DataMember] public Collection<IProjectPart> ProjectParts { get; set; } 

    public T GetPart<T>() where T : IProjectPart 
    { 
     return ProjectParts.OfType<T>().First(); 
    } 

    public void RegisterPart<T>(T part) where T : IProjectPart 
    { 
     if (ProjectParts.Any(p => p.GetType().IsInstanceOfType(part))) throw new InvalidOperationException("Part already registered."); 
     ProjectParts.Add(part); 
     part.Project = this; 
    } 

    public void Load() 
    { 
     foreach (var projectPart in ProjectParts) 
     { 
      projectPart.Load(); 
     } 
    } 

    public void Unload() 
    { 
     foreach (var projectPart in ProjectParts) 
     { 
      projectPart.Unload(); 
     } 
    } 

    public void Save() 
    { 
     foreach (var projectPart in ProjectParts) 
     { 
      projectPart.Save(); 
     } 
    } 

    public Project() 
    { 
     ProjectParts = new Collection<IProjectPart>(); 
    } 
} 

谢谢！

Answer 1

问题很简单 - 当你反复序列化你的对象时，你可以使用不同大小的IProjectPart集合。该File.Open方法不清除以前的内容文件，以便承担下列步骤：

我）序列化对象有两个IProjectPart instaces - 让我们说，这将需要10行xml文件

II）序列化对象的再一个IProjectPart集合中的实例 - 这次将需要8行xml文件

iii）第9行和第10行将填充旧的xml数据，因为它们在序列化尝试之间未被清​​除 - 所以存在一些重复垃圾 - 寻找XML数据。

亲身体验一下，您会看到如何生成这些多个标签。

注：8条10日线都是近似值为我实现

注2：我建议使用using语句序列化方法里面的流（为所有IDisposable的对象）：

private void SerializeToFile(object objectToSerialize) 
{ 
    using(var stream = File.Open(_fileName, FileMode.OpenOrCreate, FileAccess.Write)) 
    { 
     using (var writer = XmlWriter.Create(stream, new XmlWriterSettings { Indent = true })) 
     { 
      NetDataContractSerializer serializer = new NetDataContractSerializer(); 
      serializer.WriteObject(writer, objectToSerialize); 
     } 
    } 
}