前言
我在“浅谈 ZipInteger”一文中的 ZipInteger 结构中使用了 BitConverter 类的 GetBytes 方法。当时我是假设 GetBytes 方法根据 IsLittleEndian 的值不同而按照不同的顺序返回字节数组。但是 MSDN 有关 BitConverter 类的文档中没有对此作出明确的说明。请参见我在 MSDN 论坛的一个贴子“请问 BitConverter.GetBytes 方法以什么顺序返回字节数组”：
在 MSDN 文档的“BigInteger 构造函数 (Byte[])”中提到： value 数组中的各个字节应该为 little-endian 顺序，从最低序位字节到最高序位字节。 将数值转换为字节数组的大多数方法，例如 BigInteger.ToByteArray 和 BitConverter.GetBytes，以 little-endian 顺序返回字节数组。 按照我的理解，该 MSDN 文档认为 BitConverter.GetBytes 方法总是以 little-endian 顺序返回字节数组，而不管 BitConverter.IsLittleEndian 的值如何。 而在 BitConverter 类的文档中并没有明确指出 BitConverter.GetBytes 方法应该以什么顺序返回字节数组。 不知我的理解是否正确。
该贴子并没有得到靠谱的答复。
测试程序
那么，我们写个程序来测试一下吧。下面就是 BitConverterTester.cs：
using System;
namespace Skyiv.Tester
{
static class BitConverterTester
{
static void Main()
{
Console.WriteLine(" OS Version: " + Environment.OSVersion);
Console.WriteLine(" CLR Version: " + Environment.Version);
Console.WriteLine(" IsLittleEndian: " + BitConverter.IsLittleEndian);
long n = 0x1234567890ABCDEF;
double d = 1;
Console.WriteLine(n.ToString("X") + ": " + BitConverter.ToString(BitConverter.GetBytes(n)));
Console.WriteLine(d.ToString("F14") + ": " + BitConverter.ToString(BitConverter.GetBytes(d)));
}
}
}
这个程序在 Windows Server 2003 操作系统的 .NET Framework 4 环境下编译和运行：
C:\CS\BitConverterTester> csc BitConverterTester.cs
Microsoft(R) Visual C# 2010 编译器 4.0.30319.1 版
版权所有(C) Microsoft Corporation。保留所有权利。
C:\CS\BitConverterTester> BitConverterTester
OS Version: Microsoft Windows NT 5.2.3790 Service Pack 2
CLR Version: 4.0.30319.1
IsLittleEndian: True
1234567890ABCDEF: EF-CD-AB-90-78-56-34-12
1.00000000000000: 00-00-00-00-00-00-F0-3F
C:\CS\BitConverterTester>
在 Ubuntu 10.10 操作系统的 Mono 2.8.2 环境下编译和运行：
ben@ben-m4000t:~/work/BitConverterTester$ dmcs BitConverterTester.cs
ben@ben-m4000t:~/work/BitConverterTester$ mono28 BitConverterTester.exe
OS Version: Unix 2.6.35.24
CLR Version: 4.0.30319.1
IsLittleEndian: True
1234567890ABCDEF: EF-CD-AB-90-78-56-34-12
1.00000000000000: 00-00-00-00-00-00-F0-3F
ben@ben-m4000t:~/work/BitConverterTester$
这两次运行的结果都在预料之中，BitConverter 类的 GetBytes 方法以 Little-Endian 顺序返回字节数组。但是，在这两次运行中，IsLittleEndian 的值都为 True，所以还是没有解决我们的问题。
查看 Microsoft .NET Framework 4 中相关的源程序代码
祭出 Reflector 这个神器：
在 mscorlib.dll 的 System 命名空间下找到 BitConverter 类:
如上图如示，IsLittleEndian 是 BitConverter 类的静态只读字段。
如上图所示，在 BitConverter 类的静态构造函数中，直接把 IsLittleEndian 这个静态只读字段的值赋值为 true。由于我没有 Microsoft 实现 BitConverter 类的 C# 源程序代码，不知道是 Microsoft 的 C# 源程序中就是直接这样写呢，还是实际上是有根据平台来判断的，但是 C# 编译器在具体平台上优化了这段代码。
如上图所示，GetBytes(Int64) 方法也非常简单，直接通过不安全的指针转换就得到了相应的字节数组。学过 C 语言中的朋友想必非常熟悉这种做法。这下清楚了，BitConverter 类的 GetBytes 方法是根据 IsLittleEndian 的值的不同按照不同顺序来返回字节数组的。MSDN 文档在“BigInteger 构造函数 (Byte[]) ”中的相关说法：“例如 BigInteger.ToByteArray 和 BitConverter.GetBytes，以 little-endian 顺序返回字节数组”是不正确的。
继续看下去：
上图中的 ToInt64 方法也明确地根据 IsLittleEndian 的值采取不同的动作。其实，既然在静态构造函数中明确地给 IsLittleEndian 赋值为 true，这里也可以省略对 IsLittleEndian 的判断，直接按照 IsLittleEndian 的值为 true 去做就行了，可以稍微节省点代码，提高点速度。
我们看到，GetBytes(Double) 方法也非常简单。
查看 Mono 2.8.2 中相关的源代码
我们可以到 http://ftp.novell.com/pub/mono/sources/mono/ 下载 Mono 的源代码，然后按以下方法找到 BitConverter 类的源程序代码：
ben@ben-m4000t:~$ cd src/mono-2.8.2 ben@ben-m4000t:~/src/mono-2.8.2$ find . -name BitConverter.cs ./mcs/class/corlib/System/BitConverter.cs ben@ben-m4000t:~/src/mono-2.8.2$
下面就是 BitConverter.cs：
//
// System.BitConverter.cs
//
// Author:
// Matt Kimball (matt@kimball.net)
//
//
// Copyright (C) 2004 Novell, Inc (http://www.novell.com)
//
// Permission is hereby granted, free of charge, to any person obtaining
// a copy of this software and associated documentation files (the
// "Software"), to deal in the Software without restriction, including
// without limitation the rights to use, copy, modify, merge, publish,
// distribute, sublicense, and/or sell copies of the Software, and to
// permit persons to whom the Software is furnished to do so, subject to
// the following conditions:
//
// The above copyright notice and this permission notice shall be
// included in all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
// EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
// MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
// LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
// OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
// WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
//
using System.Text;
namespace System
{
public
static
class BitConverter
{
static readonly bool SwappedWordsInDouble = DoubleWordsAreSwapped ();
public static readonly bool IsLittleEndian = AmILittleEndian ();
static unsafe bool AmILittleEndian ()
{
// binary representations of 1.0:
// big endian: 3f f0 00 00 00 00 00 00
// little endian: 00 00 00 00 00 00 f0 3f
// arm fpa little endian: 00 00 f0 3f 00 00 00 00
double d = 1.0;
byte *b = (byte*)&d;
return (b [0] == 0);
}
static unsafe bool DoubleWordsAreSwapped ()
{
// binary representations of 1.0:
// big endian: 3f f0 00 00 00 00 00 00
// little endian: 00 00 00 00 00 00 f0 3f
// arm fpa little endian: 00 00 f0 3f 00 00 00 00
double d = 1.0;
byte *b = (byte*)&d;
return b [2] == 0xf0;
}
public static long DoubleToInt64Bits (double value)
{
return ToInt64 (GetBytes (value), 0);
}
public static double Int64BitsToDouble (long value)
{
return ToDouble (GetBytes (value), 0);
}
internal static double InternalInt64BitsToDouble (long value)
{
return SwappableToDouble (GetBytes (value), 0);
}
unsafe static byte[] GetBytes (byte *ptr, int count)
{
byte [] ret = new byte [count];
for (int i = 0; i < count; i++) {
ret [i] = ptr [i];
}
return ret;
}
unsafe public static byte[] GetBytes (bool value)
{
return GetBytes ((byte *) &value, 1);
}
unsafe public static byte[] GetBytes (char value)
{
return GetBytes ((byte *) &value, 2);
}
unsafe public static byte[] GetBytes (short value)
{
return GetBytes ((byte *) &value, 2);
}
unsafe public static byte[] GetBytes (int value)
{
return GetBytes ((byte *) &value, 4);
}
unsafe public static byte[] GetBytes (long value)
{
return GetBytes ((byte *) &value, 8);
}
[CLSCompliant (false)]
unsafe public static byte[] GetBytes (ushort value)
{
return GetBytes ((byte *) &value, 2);
}
[CLSCompliant (false)]
unsafe public static byte[] GetBytes (uint value)
{
return GetBytes ((byte *) &value, 4);
}
[CLSCompliant (false)]
unsafe public static byte[] GetBytes (ulong value)
{
return GetBytes ((byte *) &value, 8);
}
unsafe public static byte[] GetBytes (float value)
{
return GetBytes ((byte *) &value, 4);
}
unsafe public static byte[] GetBytes (double value)
{
if (SwappedWordsInDouble) {
byte[] data = new byte [8];
byte *p = (byte*)&value;
data [0] = p [4];
data [1] = p [5];
data [2] = p [6];
data [3] = p [7];
data [4] = p [0];
data [5] = p [1];
data [6] = p [2];
data [7] = p [3];
return data;
} else {
return GetBytes ((byte *) &value, 8);
}
}
unsafe static void PutBytes (byte *dst, byte[] src, int start_index, int count)
{
if (src == null)
throw new ArgumentNullException ("value");
if (start_index < 0 || (start_index > src.Length - 1))
throw new ArgumentOutOfRangeException ("startIndex", "Index was"
+ " out of range. Must be non-negative and less than the"
+ " size of the collection.");
// avoid integer overflow (with large pos/neg start_index values)
if (src.Length - count < start_index)
throw new ArgumentException ("Destination array is not long"
+ " enough to copy all the items in the collection."
+ " Check array index and length.");
for (int i = 0; i < count; i++)
dst[i] = src[i + start_index];
}
unsafe public static bool ToBoolean (byte[] value, int startIndex)
{
if (value == null)
throw new ArgumentNullException ("value");
if (startIndex < 0 || (startIndex > value.Length - 1))
throw new ArgumentOutOfRangeException ("startIndex", "Index was"
+ " out of range. Must be non-negative and less than the"
+ " size of the collection.");
if (value [startIndex] != 0)
return true;
return false;
}
unsafe public static char ToChar (byte[] value, int startIndex)
{
char ret;
PutBytes ((byte *) &ret, value, startIndex, 2);
return ret;
}
unsafe public static short ToInt16 (byte[] value, int startIndex)
{
short ret;
PutBytes ((byte *) &ret, value, startIndex, 2);
return ret;
}
unsafe public static int ToInt32 (byte[] value, int startIndex)
{
int ret;
PutBytes ((byte *) &ret, value, startIndex, 4);
return ret;
}
unsafe public static long ToInt64 (byte[] value, int startIndex)
{
long ret;
PutBytes ((byte *) &ret, value, startIndex, 8);
return ret;
}
[CLSCompliant (false)]
unsafe public static ushort ToUInt16 (byte[] value, int startIndex)
{
ushort ret;
PutBytes ((byte *) &ret, value, startIndex, 2);
return ret;
}
[CLSCompliant (false)]
unsafe public static uint ToUInt32 (byte[] value, int startIndex)
{
uint ret;
PutBytes ((byte *) &ret, value, startIndex, 4);
return ret;
}
[CLSCompliant (false)]
unsafe public static ulong ToUInt64 (byte[] value, int startIndex)
{
ulong ret;
PutBytes ((byte *) &ret, value, startIndex, 8);
return ret;
}
unsafe public static float ToSingle (byte[] value, int startIndex)
{
float ret;
PutBytes ((byte *) &ret, value, startIndex, 4);
return ret;
}
unsafe public static double ToDouble (byte[] value, int startIndex)
{
double ret;
if (SwappedWordsInDouble) {
byte* p = (byte*)&ret;
if (value == null)
throw new ArgumentNullException ("value");
if (startIndex < 0 || (startIndex > value.Length - 1))
throw new ArgumentOutOfRangeException ("startIndex", "Index was"
+ " out of range. Must be non-negative and less than the"
+ " size of the collection.");
// avoid integer overflow (with large pos/neg start_index values)
if (value.Length - 8 < startIndex)
throw new ArgumentException ("Destination array is not long"
+ " enough to copy all the items in the collection."
+ " Check array index and length.");
p [0] = value [startIndex + 4];
p [1] = value [startIndex + 5];
p [2] = value [startIndex + 6];
p [3] = value [startIndex + 7];
p [4] = value [startIndex + 0];
p [5] = value [startIndex + 1];
p [6] = value [startIndex + 2];
p [7] = value [startIndex + 3];
return ret;
}
PutBytes ((byte *) &ret, value, startIndex, 8);
return ret;
}
unsafe internal static double SwappableToDouble (byte[] value, int startIndex)
{
double ret;
if (SwappedWordsInDouble) {
byte* p = (byte*)&ret;
if (value == null)
throw new ArgumentNullException ("value");
if (startIndex < 0 || (startIndex > value.Length - 1))
throw new ArgumentOutOfRangeException ("startIndex", "Index was"
+ " out of range. Must be non-negative and less than the"
+ " size of the collection.");
// avoid integer overflow (with large pos/neg start_index values)
if (value.Length - 8 < startIndex)
throw new ArgumentException ("Destination array is not long"
+ " enough to copy all the items in the collection."
+ " Check array index and length.");
p [0] = value [startIndex + 4];
p [1] = value [startIndex + 5];
p [2] = value [startIndex + 6];
p [3] = value [startIndex + 7];
p [4] = value [startIndex + 0];
p [5] = value [startIndex + 1];
p [6] = value [startIndex + 2];
p [7] = value [startIndex + 3];
return ret;
} else if (!IsLittleEndian) {
byte* p = (byte*)&ret;
if (value == null)
throw new ArgumentNullException ("value");
if (startIndex < 0 || (startIndex > value.Length - 1))
throw new ArgumentOutOfRangeException ("startIndex", "Index was"
+ " out of range. Must be non-negative and less than the"
+ " size of the collection.");
// avoid integer overflow (with large pos/neg start_index values)
if (value.Length - 8 < startIndex)
throw new ArgumentException ("Destination array is not long"
+ " enough to copy all the items in the collection."
+ " Check array index and length.");
p [0] = value [startIndex + 7];
p [1] = value [startIndex + 6];
p [2] = value [startIndex + 5];
p [3] = value [startIndex + 4];
p [4] = value [startIndex + 3];
p [5] = value [startIndex + 2];
p [6] = value [startIndex + 1];
p [7] = value [startIndex + 0];
return ret;
}
PutBytes ((byte *) &ret, value, startIndex, 8);
return ret;
}
public static string ToString (byte[] value)
{
if (value == null)
throw new ArgumentNullException ("value");
return ToString (value, 0, value.Length);
}
public static string ToString (byte[] value, int startIndex)
{
if (value == null)
throw new ArgumentNullException ("value");
return ToString (value, startIndex, value.Length - startIndex);
}
public static string ToString (byte[] value, int startIndex, int length)
{
if (value == null)
throw new ArgumentNullException ("byteArray");
// The 4th and last clause (start_index >= value.Length)
// was added as a small fix to a very obscure bug.
// It makes a small difference when start_index is
// outside the range and length==0.
if (startIndex < 0 || startIndex >= value.Length) {
// special (but valid) case (e.g. new byte [0])
if ((startIndex == 0) && (value.Length == 0))
return String.Empty;
throw new ArgumentOutOfRangeException ("startIndex", "Index was"
+ " out of range. Must be non-negative and less than the"
+ " size of the collection.");
}
if (length < 0)
throw new ArgumentOutOfRangeException ("length",
"Value must be positive.");
// note: re-ordered to avoid possible integer overflow
if (startIndex > value.Length - length)
throw new ArgumentException ("startIndex + length > value.Length");
if (length == 0)
return string.Empty;
StringBuilder builder = new StringBuilder(length * 3 - 1);
int end = startIndex + length;
for (int i = startIndex; i < end; i++) {
if (i > startIndex)
builder.Append('-');
char high = (char)((value[i] >> 4) & 0x0f);
char low = (char)(value[i] & 0x0f);
if (high < 10)
high += '0';
else {
high -= (char) 10;
high += 'A';
}
if (low < 10)
low += '0';
else {
low -= (char) 10;
low += 'A';
}
builder.Append(high);
builder.Append(low);
}
return builder.ToString ();
}
}
}
可以看出，在第 39 行通过第 41 到 50 行的 AmILittleEndian 方法给 IsLittleEndian 这个静态只读字段赋值。这个 AmILittleEndian 方法中的注释写得很明白，其实有三种不台的平台，除了 Big-Endian 和 Little-Endian 外，还有一种“arm fpa little endian”，在这里也归于 Little-Endian。但是在后面的方法中是要以不同的手段进行处理的。
另外，不象 Microsoft .NET Framework 4，Mono 中 BitConverter 类的各种重载的公有静态GetBytes 方法是统一调用第 78 到 87 行的私有静态 GetBytes 方法来实现其功能的。
从 Mono 中相关的源程序代码中也可以看出，BitConverter 类的 GetBytes 方法是根据 IsLittleEndian 的值的不同以不同的顺序返回字节数组的。
结论
总结一下，MSDN 文档在“BigInteger 构造函数 (Byte[]) ”中的相关说法：“例如 BigInteger.ToByteArray 和 BitConverter.GetBytes，以 little-endian 顺序返回字节数组”是不正确的。BitConverter 类的 GetBytes 方法是根据 IsLittleEndian 的值的不同以不同的顺序返回字节数组的。
