当 xml 有 indentation/line 中断时，字典的自定义序列化失败

Question

为了使字典序列化更清晰 XML，我编写了一个自定义 class，它实现了 IXmlSerializable.

我的自定义class是这样定义的：

public class MyCollection : System.Collections.Generic.Dictionary<string, string>, IXmlSerializable
{
    private const string XmlElementName = "MyData";
    private const string XmlAttributeId = "Id";

    public XmlSchema GetSchema()
    {
        return null;
    }

    public void ReadXml(XmlReader reader)
    {
        reader.MoveToContent();
        while (reader.Read())
        {
            if(reader.LocalName == XmlElementName)
            {
                var tag = reader.GetAttribute(XmlAttributeId);
                var content = reader.ReadElementContentAsString(); 

                this.Add(tag, content);
            }
        }
    }

    public void WriteXml(System.Xml.XmlWriter writer)
    {
        foreach (string key in this.Keys)
        {
            writer.WriteStartElement(XmlElementName);
            writer.WriteAttributeString(XmlAttributeId, key);
            writer.WriteString(this[key]);
            writer.WriteEndElement();
        }
    }
}

我的代码适用于这个 XML 片段：

<MyCollection xmlns="http://schemas.datacontract.org/2004/07/MyProject" xmlns:i="http://www.w3.org/2001/XMLSchema-instance">
    <MyData Id="1">some content</MyData>
    <MyData Id="2">some other content</MyData>
</MyCollection>

但是，当我缩小 XML 时，我的代码抛出异常：

<MyCollection xmlns="http://schemas.datacontract.org/2004/07/MyProject" xmlns:i="http://www.w3.org/2001/XMLSchema-instance"><MyData Id="1">some content </MyData><MyData Id="2">some other content</MyData></MyCollection>

例外情况是：

System.InvalidOperationException: The ReadElementContentAsString method is not supported on node type EndElement

它是在调用 ReadElementContentAsString 时抛出的。

如何修复我的代码？

我可以使用 :

重现问题

var xml = @"<MyCollection xmlns=""http://schemas.datacontract.org/2004/07/MyProject"" xmlns:i=""http://www.w3.org/2001/XMLSchema-instance""><MyData Id=""1"">some content </MyData><MyData Id=""2"">some other content</MyData></MyCollection>";

var raw = Encoding.UTF8.GetBytes(xml);

var serializer = new DataContractSerializer(typeof(MyCollection));

using (var ms = new MemoryStream(raw))
{
    var result = serializer.ReadObject(ms); // Exception throws here
}

Answer 1

您的问题是 reader.ReadElementContentAsString() positions the reader at the beginning of the next node, not the end of the current node. Then, your subsequent unconditional call to reader.Read() 消耗了下一个节点。当该节点是空白时，不会造成任何伤害，但当该节点是一个元素时，该元素将被跳过。

您的 MyCollection 的以下版本修复了此问题：

public class MyCollection : System.Collections.Generic.Dictionary<string, string>, IXmlSerializable
{
    public XmlSchema GetSchema()
    {
        return null;
    }

    public void ReadXml(XmlReader reader)
    {
        using (var subReader = reader.ReadSubtree())
        {
            XmlKeyValueListHelper.ReadKeyValueXml(subReader, this);
        }
        // Consume the EndElement also (or move past the current element if reader.IsEmptyElement).
        reader.Read();
    }

    public void WriteXml(System.Xml.XmlWriter writer)
    {
        XmlKeyValueListHelper.WriteKeyValueXml(writer, this);
    }
}

public static class XmlKeyValueListHelper
{
    private const string XmlElementName = "MyData";
    private const string XmlAttributeId = "Id";

    public static void WriteKeyValueXml(System.Xml.XmlWriter writer, ICollection<KeyValuePair<string, string>> collection)
    {
        foreach (var pair in collection)
        {
            writer.WriteStartElement(XmlElementName);
            writer.WriteAttributeString(XmlAttributeId, pair.Key);
            writer.WriteString(pair.Value);
            writer.WriteEndElement();
        }
    }

    public static void ReadKeyValueXml(System.Xml.XmlReader reader, ICollection<KeyValuePair<string, string>> collection)
    {
        if (reader.IsEmptyElement)
        {
            reader.Read();
            return;
        }

        reader.ReadStartElement(); // Advance to the first sub element of the list element.
        while (reader.NodeType != XmlNodeType.EndElement)
        {
            if (reader.NodeType == XmlNodeType.Element && reader.LocalName == XmlElementName)
            {
                var tag = reader.GetAttribute(XmlAttributeId);
                string content;
                if (reader.IsEmptyElement)
                {
                    content = string.Empty;
                    // Move past the end of item element
                    reader.Read();
                }
                else
                {
                    // Read content and move past the end of item element
                    content = reader.ReadElementContentAsString();
                }
                collection.Add(new KeyValuePair<string, string>(tag, content));
            }
            else
            {
                // For instance a comment.
                reader.Skip();
            }
        }
        // Move past the end of the list element
        reader.ReadEndElement();
    }
}

一些注意事项：

通过使用 XmlReader.ReadSubtree()，我确保 ReadXml() 不会读取超过 MyCollection 元素的末尾，从而破坏未来的元素——这是一个容易犯的错误实施 IXmlSerializable.
通过检查 reader.NodeType == XmlNodeType.Element && reader.LocalName == XmlElementName 我忽略了意外类型的节点，例如评论。

工作.Net fiddle。

当 xml 有 indentation/line 中断时，字典的自定义序列化失败

Custom serialization of a dictionary fails when xml has indentation/line break

c#

xml

serialization

ixmlserializable