检测字节流是否是UTF8编码

几天前偶尔看到有人发帖子问“如何自动识别判断url中的中文参数是GB2312还是Utf-8编码”

也拜读了wcwtitxu使用巨牛的正则表达式检测UTF8编码的算法。

使用无数或条件的正则表达式用起来却是性能不高。

刚好曾经在项目中有类似的需求，这里把处理思路和整理后的源代码贴出来供大家参考

先聊聊原理：

UTF8的编码规则如下表

看起来很复杂，总结起来如下：

ASCII码（U+0000 - U+007F），不编码

其余编码规则为

•第一个Byte二进制以形式为n个1紧跟个0 (n >= 2), 0后面的位数用来存储真正的字符编码，n的个数说明了这个多Byte字节组字节数（包括第一个Byte）
•结下来会有n个以10开头的Byte，后6个bit存储真正的字符编码。
因此对整个编码byte流进行分析可以得出是否是UTF8编码的判断。

根据这个规则，我给出的C#代码如下：

        /// <summary>
        ///   Determines whether the given <paramref name="inputStream"/>is UTF8 encoding bytes.
        /// </summary>
        /// <param name="inputStream">
        ///    The input stream.
        ///  </param>
        /// <returns>
        ///   <see langword="true"/> if given bystes stream is in UTF8 encoding; otherwise, <see langword="false"/>.
        /// </returns>
        /// <remarks>
        ///   All ASCII chars will regards not UTF8 encoding.
        /// </remarks>
        public static bool IsTextUTF8(ref byte[] inputStream)
        {
            int encodingBytesCount = 0;
            bool allTextsAreASCIIChars = true;

            for (int i = 0; i < inputStream.Length; i++)
            {
                byte current = inputStream[i];

                if ((current & 0x80) == 0x80)
                {                    
                    allTextsAreASCIIChars = false;
                }
                // First byte
                if (encodingBytesCount == 0)
                {
                    if ((current & 0x80) == 0)
                    {
                        // ASCII chars, from 0x00-0x7F
                        continue;
                    }

                    if ((current & 0xC0) == 0xC0)
                    {
                        encodingBytesCount = 1;
                        current <<= 2;

                        // More than two bytes used to encoding a unicode char.
                        // Calculate the real length.
                        while ((current & 0x80) == 0x80)
                        {
                            current <<= 1;
                            encodingBytesCount++;
                        }
                    }                    
                    else
                    {
                        // Invalid bits structure for UTF8 encoding rule.
                        return false;
                    }
                }                
                else
                {
                    // Following bytes, must start with 10.
                    if ((current & 0xC0) == 0x80)
                    {                        
                        encodingBytesCount--;
                    }
                    else
                    {
                        // Invalid bits structure for UTF8 encoding rule.
                        return false;
                    }
                }
            }

            if (encodingBytesCount != 0)
            {
                // Invalid bits structure for UTF8 encoding rule.
                // Wrong following bytes count.
                return false;
            }

            // Although UTF8 supports encoding for ASCII chars, we regard as a input stream, whose contents are all ASCII as default encoding.
            return !allTextsAreASCIIChars;
        }

再附上单元测试代码：

    /// <summary>
    ///This is a test class for EncodingHelperTest and is intended
    ///to contain all EncodingHelperTest Unit Tests
    ///</summary>
    [TestClass()]
    public class EncodingHelperTest
    {
        /// <summary>
        ///  Normal test for this method.
        ///</summary>
        [TestMethod()]
        public void IsTextUTF8Test()
        {
            for (int i = 0; i < 1000; i++)
            {
                List<Char> chars = new List<char>();
                chars.Add('中');

                List<UnicodeCategory> temp = new List<UnicodeCategory>();
                Random rd = new Random((int)(DateTime.Now.Ticks & 0x7FFFFFFF));

                for (int j = 0; j < 255; j++)
                {
                    char ch = (char)rd.Next(0xFFFF);
                    UnicodeCategory uc = System.Globalization.CharUnicodeInfo.GetUnicodeCategory(ch);
                    if (uc == UnicodeCategory.Surrogate || // Single surrogate could not be encoding correctly.
                        uc == UnicodeCategory.PrivateUse || // Private use blocks should be excluded.
                        uc == UnicodeCategory.OtherNotAssigned
                        )
                    {
                        j--;
                    }
                    else
                    {
                        chars.Add(ch);
                        temp.Add(uc);
                    }
                }

                string str = new string(chars.ToArray());

                byte[] inputStream = Encoding.UTF8.GetBytes(str);
                bool expected = true; 
                bool actual;
                actual = EncodingHelper.IsTextUTF8(ref inputStream);
                Assert.AreEqual(expected, actual, string.Format("UTF8_Assert Fails at:{0}", str));

                inputStream = Encoding.GetEncoding(932).GetBytes(str);
                expected = false;

                actual = EncodingHelper.IsTextUTF8(ref inputStream);
                Assert.AreEqual(expected, actual, string.Format("ShiftJIS_Assert Fails at:{0}", str));
            }
        }

        /// <summary>
        ///   Check with All ASCII chars
        /// </summary>
        [TestMethod]
        public void IsTextUTF8Test_AllASCII()
        {
            string str = "ABCDEFGHKLHSJKLDFHJKLHAJKLSHJKLHAJKLSHDJKLAHSDJKLHAJKLSDHJKLASHDJKLHASJKLDHJKLASD";

            byte[] inputStream = Encoding.UTF8.GetBytes(str);
            bool expected = false;
            bool actual;
            actual = EncodingHelper.IsTextUTF8(ref inputStream);
            Assert.AreEqual(expected, actual, string.Format("UTF8_Assert Fails at:{0}", str));


        }
    }

另：

如果是判断一个文件是否使用了UTF8编码，不一定非用这种方法，因为通常以UTF8格式保存的文件最初两个字符是BOM头，标示该文件使用了UTF8编码。

参考：

维基百科：http://en.wikipedia.org/wiki/UTF-8