尝试将结构转换为 byte[] 时的方法访问期望

本文关键字：方法期望访问 byte 结构转换 | 更新日期: 2023-09-27 18:35:52

我需要将以下结构转换为字节数组：

[Serializable]
public struct newLeads
{
    public string id;
    public string first_name;
    public string last_name;
}

我正在尝试使用以下代码转换为字节数组：

public class ConvertStruct
{
    public static byte[] StructureToByteArray(object obj)
    {
        int Length = Marshal.SizeOf(obj);
        byte[] bytearray = new byte[Length];
        IntPtr ptr = Marshal.AllocHGlobal(Length);
        Marshal.StructureToPtr(obj, ptr, false);
        Marshal.Copy(ptr, bytearray, 0, Length);
        Marshal.FreeHGlobal(ptr);
        return bytearray;
    }
}

我在网上收到异常：

IntPtr ptr = Marshal.AllocHGlobal(Length);

异常：安全透明方法"Classes.ConvertStruct.ConvertStruct.StructureToByteArray（System.Object）"访问安全关键方法"System.Runtime.InteropServices.Marshal.AllocHGlobal（Int32）"失败。

我的问题是？如何解决此问题以避免异常并将我的简单结构转换为byte[]？

提前感谢！

更新：我在控制台应用程序中尝试过这个，它可以工作。我从一个 asp.net 页面代码隐藏中调用它，所以这肯定与它有关，但我无法弄清楚是什么！

尝试将结构转换为 byte[] 时的方法访问期望

检查结构的大小，不要将其框起来。

为字符串设置一个适当的封送处理（从 goric 的答案中获取），也就是说，你会得到的是字节数组中字符串的内存地址（不是一件好事）。

[Serializable]
public struct newLeads
{
    [MarshalAs(UnmanagedType.ByValTStr, SizeConst = 5000)]
    public string id;
    [MarshalAs(UnmanagedType.ByValTStr, SizeConst = 5000)]
    public string first_name;
    [MarshalAs(UnmanagedType.ByValTStr, SizeConst = 5000)]
    public string last_name;
}
[SecurityPermission(SecurityAction.Demand, UnmanagedCode = true)]
public static byte[] ToByteArray(newLeads value)
{
    int length = Marshal.SizeOf(typeof(newLeads));
    var result = new byte[length];
    IntPtr sourcePtr = Marshal.AllocHGlobal(length);
    Marshal.StructureToPtr(value, sourcePtr, false);
    Marshal.Copy(sourcePtr, result, 0, length);
    Marshal.FreeHGlobal(sourcePtr);
    return result;
}

在您的评论中，您说此代码失败得更快。好吧，它使用安全权限要求（如 .NET 4 建议的那样），每次调用该方法时，它都会检查特定的性能。您可以尝试在没有它的情况下执行它，预期的结果就是您一开始得到的结果。

真正的答案

您必须在受约束的环境中运行，可能是某种不支持指针的沙盒或平台。在这种情况下，我们可能需要通过其他方式进行转换。

你说 ASP.NET？就是这样。

为了在没有指针的情况下进行转换，请尝试以下技术：

//You will have to decide an encoding.
//If nto ASCII, try UTF8Encoding, UnicodeEncoding or (Hopefully not) UTF7Encoding
void Main()
{
    Encoding encoding = new ASCIIEncoding();
    newLeads target = GetNewLeads();
    byte[] id = EncodeString(target.id, encoding);
    byte[] first_name = EncodeString(target.first_name, encoding);
    byte[] last_name = EncodeString(target.last_name, encoding);
}
byte[] EncodeString(string str, Encoding encoding)
{
    byte[] data;
    if (ReferenceEquals(str, null))
    {
        data = new byte[0];
    }
    else
    {
        data = encoding.GetBytes(str);
    }
    return data;
}

在这一点上，我需要更多地了解您的情况，以便为您提供更好的解决方案，特别是谁或什么将读取该字节数组？无论如何，您可以对字符串的长度进行编码，如下所示（为 null 保留 -1）：

byte[] EncodeString(string str, Encoding encoding)
{
    byte[] data;
    byte[] data_length;
    Union union = new Union();
    if (ReferenceEquals(str, null))
    {
        data = new byte[0];
        union.data = -1;
    }
    else
    {
        data = encoding.GetBytes(str);
        union.data = str.Length;
    }
    data_length = new byte[]{union.a, union.b, union.c, union.c};
    int length = data.Length;
    byte[] result = new byte[4 + data.Length];
    System.Buffer.BlockCopy(data_length, 0, result, 0, 4);
    System.Buffer.BlockCopy(data, 0, result, 4, length);
    return result;
}
//I hope endianess doesn't bite
[StructLayout(LayoutKind.Explicit)] 
struct Union
{
    [FieldOffset(0)]
    public int data;
    [FieldOffset(0)] 
    public byte a;
    [FieldOffset(1)] 
    public byte b;
    [FieldOffset(2)] 
    public byte c;
    [FieldOffset(3)] 
    public byte d;
}

最后，我们需要连接这些数组。我知道这可以进一步优化...（使用MemoryStream和StreamWriter的好主意）无论如何，这是我的第一个实现[测试]：

byte[] ToByteArray(newLeads value)
{
    Encoding encoding = new ASCIIEncoding(); //Choose some encoding
    byte[] id = EncodeString(value.id, encoding);
    byte[] first_name = EncodeString(value.first_name, encoding);
    byte[] last_name = EncodeString(value.last_name, encoding);
    byte[] result = new byte[id.Length + first_name.Length + last_name.Length];
    System.Buffer.BlockCopy(id, 0, result, 0, id.Length);
    System.Buffer.BlockCopy
        (
            first_name,
            0,
            result,
            id.Length,
            first_name.Length
        );
    System.Buffer.BlockCopy
        (
            last_name,
            0,
            result,
            id.Length + first_name.Length,
            last_name.Length
        );
    return result;
}
byte[] EncodeString(string str, Encoding encoding)
{
    byte[] data;
    byte[] data_length;
    Union union = new Union();
    if (ReferenceEquals(str, null))
    {
        data = new byte[0];
        union.data = -1;
    }
    else
    {
        data = encoding.GetBytes(str);
        union.data = str.Length;
    }
    data_length = new byte[]{union.a, union.b, union.c, union.c};
    int length = data.Length;
    byte[] result = new byte[4 + data.Length];
    System.Buffer.BlockCopy(data_length, 0, result, 0, 4);
    System.Buffer.BlockCopy(data, 0, result, 4, length);
    return result;
}
//I hope endianess doesn't bite
[StructLayout(LayoutKind.Explicit)] 
struct Union
{
    [FieldOffset(0)]
    public int data;
    [FieldOffset(0)] 
    public byte a;
    [FieldOffset(1)] 
    public byte b;
    [FieldOffset(2)] 
    public byte c;
    [FieldOffset(3)] 
    public byte d;
}

注意：我没有使用以空结尾的字符串，因为我不知道您最终会使用哪种编码。

优化

相同的逻辑，但用流实现（联合没有改变）[测试]。

byte[] ToByteArray(newLeads value)
{
    Encoding encoding = new ASCIIEncoding(); //Choose some encoding
    var stream = new MemoryStream();
    EncodeString(value.id, stream, encoding);
    EncodeString(value.first_name, stream, encoding);
    EncodeString(value.last_name, stream, encoding);
    int length = (int)stream.Length;
    byte[] result = new byte[(int)stream.Length];
    System.Buffer.BlockCopy(stream.GetBuffer(), 0, result, 0, length);
    stream.Close();
    return result;
}
void EncodeString(string str, Stream stream, Encoding encoding)
{
    Union union = new Union();
    if (ReferenceEquals(str, null))
    {
        union.data = -1;
        stream.Write(new byte[]{union.a, union.b, union.c, union.c}, 0, 4);
    }
    else
    {
        union.data = str.Length;
        stream.Write(new byte[]{union.a, union.b, union.c, union.c}, 0, 4);
        var tmp = encoding.GetBytes(str);
        stream.Write(tmp, 0, tmp.Length);
    }
}

将琴弦重新放回原处

为了取回数据，我们首先读取字符串的长度（具有相同类型的 Union）：

var newLeads = GetNewLeads();
var z= ToByteArray(newLeads); //we have the byteArray in z
var data = new MemoryStream(z); //Create an stream for convenience
//Use the union to get the length
var union = new Union()
    {
        a = (byte)data.ReadByte(),
        b = (byte)data.ReadByte(),
        c = (byte)data.ReadByte(),
        d = (byte)data.ReadByte()
    };
Console.WriteLine(union.data); //the length of the first string

我们的下一步是读取这么多字符，为此我们将使用 StreamReader：

var newLeads = GetNewLeads();
var z = ToByteArray(newLeads);
var data = new MemoryStream(z);
var union = new Union()
    {
        a = (byte)data.ReadByte(),
        b = (byte)data.ReadByte(),
        c = (byte)data.ReadByte(),
        d = (byte)data.ReadByte()
    };
var encoding = new ASCIIEncoding();
string result = null;
if (union.data != -1)
{
    char[] finalChars = new char[union.data];
    var reader = new StreamReader(data, encoding);
    reader.Read(finalChars, 0, union.data);
    result = new string(finalChars);
}
Console.WriteLine(result);

有了这个，我们构建了一个解码字符串的方法：

string DecodeString(Stream data, Encoding encoding)
{   
    //TODO: You may want to validate that data and encoding are not null
    //or make this private
    var union = new Union()
    {
        a = (byte)data.ReadByte(),
        b = (byte)data.ReadByte(),
        c = (byte)data.ReadByte(),
        d = (byte)data.ReadByte()
    };
    string result = null;
    if (union.data != -1)
    {
        char[] finalChars = new char[union.data];
        var reader = new StreamReader(data, encoding);
        reader.Read(finalChars, 0, union.data);
        result = new string(finalChars);
    }
    return result;
}
//Convenience method, not needed:
string DecodeString(byte[] data, Encoding encoding)
{
    //TODO: You may want to validate that data and encoding are not null
    //or make this private
    return DecodeString(new MemoryStream(data), encoding);
}

最后是恢复新潜在客户的方法（再次具有相同的联合类型）[测试]：

newLeads FromByteArray(byte[] data)
{
    //TODO: Validate that data is not null
    Encoding encoding = new ASCIIEncoding(); //Choose the same encoding
    newLeads result = new newLeads();
    var reader = new StreamReader(new MemoryStream(data), encoding);
    result.id = DecodeString(reader);
    result.first_name = DecodeString(reader);
    result.last_name = DecodeString(reader);
    reader.Close();
    return result;
}

//Changed to reuse StreamReader...
//Because closing it will close the underlying stream
string DecodeString(StreamReader reader)
{
    //TODO: You may want to validate that reader is not null
    //or make this private
    var data = reader.BaseStream;
    var union = new Union()
    {
        a = (byte)data.ReadByte(),
        b = (byte)data.ReadByte(),
        c = (byte)data.ReadByte(),
        d = (byte)data.ReadByte()
    };
    string result = null;
    if (union.data != -1)
    {
        char[] finalChars = new char[union.data];
        reader.Read(finalChars, 0, union.data);
        result = new string(finalChars);
    }
    return result;
}
//Convenience method, not needed:
string DecodeString(byte[] data, Encoding encoding)
{
    //TODO: You may want to validate that data and encoding are not null
    //or make this private
    return DecodeString(new StreamReader(new MemoryStream(data), encoding));
}

在结构中，使用正确的MarshalAs属性标记字符串（MSDN 参考）：

[Serializable]
public struct newLeads
{
    [MarshalAs(UnmanagedType.ByValTStr, SizeConst = 5000)]
    public string id;
    [MarshalAs(UnmanagedType.ByValTStr, SizeConst = 5000)]
    public string first_name;
    [MarshalAs(UnmanagedType.ByValTStr, SizeConst = 5000)]
    public string last_name;
}

确保SizeConst值足够大，以覆盖要使用的数据的最大长度。

看起来您可以将方法标记为SecurityCritical

http://msdn.microsoft.com/en-us/library/bb264475.aspx