C# UTF-8字符集长度获取及判断二进制块是否UTF-8编码字符集算法的实现

最新推荐文章于 2024-07-09 11:49:28 发布

liulilittle

最新推荐文章于 2024-07-09 11:49:28 发布

阅读量313

点赞数

分类专栏： C# 文章标签：算法 c# java

本文链接：https://blog.csdn.net/liulilittle/article/details/127698290

版权

C# 专栏收录该内容

94 篇文章 2 订阅

订阅专栏

        [DllImport("kernel32.dll", CallingConvention = CallingConvention.StdCall)]
        private extern static bool IsBadReadPtr(void* lp, uint ucb);

        private const byte kFirstBitMask = 0x80; // 1000000
        private const byte kSecondBitMask = 0x40; // 0100000
        private const byte kThirdBitMask = 0x20; // 0010000
        private const byte kFourthBitMask = 0x10; // 0001000
        private const byte kFifthBitMask = 0x08; // 0000100

        public static int GetUtf8Alignment(byte character)
        {
            int alignment = 1;
            if ((character & kFirstBitMask) > 0) // This means the first byte has a value greater than 127, and so is beyond the ASCII range.
            {
                if ((character & kThirdBitMask) > 0) // This means that the first byte has a value greater than 224, and so it must be at least a three-octet code point.
                {
                    if ((character & kFourthBitMask) > 0) // This means that the first byte has a value greater than 240, and so it must be a four-octet code point.
                    {
                        alignment = 4;
                    }
                    else
                    {
                        alignment = 3;
                    }
                }
                else
                {
                    alignment = 2;
                }
            }
            return alignment;
        }


        public static int GetUtf8BufferCount(byte* s)
        {
            if (s == null)
            {
                return 0;
            }
            byte* i = s;
            while (!IsBadReadPtr(i, 1))
            {
                int alignment = GetUtf8Alignment(*i);
                int character = 0;
                if (alignment == 1)
                {
                    character = *i++;
                }
                if (alignment == 2)
                {
                    character = *(short*)i;
                    i += 2;
                }
                else if (alignment == 3)
                {
                    character = (*i++ | *i++ << 8 | *i++ << 16);
                }
                else if (alignment == 4)
                {
                    character = *(int*)i;
                    i += 4;
                }
                if (character == 0)
                {
                    int len = unchecked((int)(i - (s + 1)));
                    return len < 0 ? 0 : len;
                }
            }
            return 0;
        }


        public static bool IsUTF8Buffer(byte[] buffer)
        {
            if (buffer == null)
            {
                return false;
            }
            fixed (byte* pinned = buffer)
            {
                return IsUTF8Buffer(pinned, buffer.Length);
            }
        }

        public static bool IsUTF8Buffer(byte* buffer, int count)
        {
            if (buffer == null || count <= 0)
            {
                return false;
            }
            int counter = 1;
            byte key = 0;
            for (int i = 0; i < count; i++)
            {
                key = buffer[i];
                if (counter == 1)
                {
                    if (key >= 0x80)
                    {
                        while (((key <<= 1) & 0x80) != 0)
                        {
                            counter++;
                        }
                        if (counter == 1 || counter > 6)
                        {
                            return false;
                        }
                    }
                }
                else
                {
                    if ((key & 0xC0) != 0x80)
                    {
                        return false;
                    }
                    counter--;
                }
            }
            return !(counter > 1);
        }