postgresql之base64编解码

一、简介

1.1 何为64？

将数据编码映射到[0-9][a-z][A-Z]+/, 10+ 26+26+2 = 64个字符

1.2 如何编码

• 因为只有64个字符，映射使用的索引只能有6bit [0,63]进行表示。
• 1byte = 8bit, 3*8 = 4*6, 因此3字节映射成了4字节大小
• 每6bit为单位，高位补0
• 不够4字节时，使用=进行填充

11111111 11111111 11111111
00111111 00111111 00111111 00111111

1.3 如何解码

• 编码后的数据就是映射后的字符流，现在需要反映射回原始数据。
• 字符反映射回下标
• 下标值去掉最高两位，将所有值拼接起来

二、代码实现

2.1 头文件

src/include/common/base64.h

#ifndef BASE64_H
#define BASE64_H

/* base 64 */
extern int	pg_b64_encode(const char *src, int len, char *dst, int dstlen);
extern int	pg_b64_decode(const char *src, int len, char *dst, int dstlen);
extern int	pg_b64_enc_len(int srclen);
extern int	pg_b64_dec_len(int srclen);

#endif							/* BASE64_H */

2.2 评估编解码需要的空间

根据需要多少空间，提前分配好空间

• 避免空间不足导致数据被截断，导致无法编解码
• 避免为了正确编解码而进行多次动态扩容数据存储空间

编码： 3字节 =》4字节
解码： 4字节 =》3字节

/*
 * pg_b64_enc_len
 *
 * Returns to caller the length of the string if it were encoded with
 * base64 based on the length provided by caller.  This is useful to
 * estimate how large a buffer allocation needs to be done before doing
 * the actual encoding.
 */
int
pg_b64_enc_len(int srclen)
{
	/* 3 bytes will be converted to 4 */
	return (srclen + 2) * 4 / 3;
}

/*
 * pg_b64_dec_len
 *
 * Returns to caller the length of the string if it were to be decoded
 * with base64, based on the length given by caller.  This is useful to
 * estimate how large a buffer allocation needs to be done before doing
 * the actual decoding.
 */
int
pg_b64_dec_len(int srclen)
{
	return (srclen * 3) >> 2;
}

2.3 编码

设计妙处1：

• 3byte处理一次
• pos = 2, 每次循环pos–
• pos小于0时，刚好循环三次，则有三个字节
• pos<<3, 第一次左移2<<3(16), 第二次左移1<<3(8), 第三次左移0<<3(0)
• buf为uint32
• buf的低三个字节则存储了需要处理的三个字节
• 依次从buf中取出6bit值作为下标值，从_base64映射出编码后的字符

设计妙处2：

• 当pos不等于2,则说明还有字节需要处理，但是又不够三个字节
• 至少有一个字节，则需要两个字节进行映射，因此没有判断，直接取了两次值
• 对于第三个字节，如果pos==0则说明有两个字节，直接取值，否则使用=进行填充

/*
 * BASE64
 */

static const char _base64[] =
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";

/*
 * pg_b64_encode
 *
 * Encode into base64 the given string.  Returns the length of the encoded
 * string, and -1 in the event of an error with the result buffer zeroed
 * for safety.
 */
int
pg_b64_encode(const char *src, int len, char *dst, int dstlen)
{
	char	   *p;
	const char *s,
			   *end = src + len;
	int			pos = 2;
	uint32		buf = 0;

	s = src;
	p = dst;

	while (s < end)
	{
		buf |= (unsigned char) *s << (pos << 3);
		pos--;
		s++;

		/* write it out */
		if (pos < 0)
		{
			/*
			 * Leave if there is an overflow in the area allocated for the
			 * encoded string.
			 */
			if ((p - dst + 4) > dstlen)
				goto error;

			*p++ = _base64[(buf >> 18) & 0x3f];
			*p++ = _base64[(buf >> 12) & 0x3f];
			*p++ = _base64[(buf >> 6) & 0x3f];
			*p++ = _base64[buf & 0x3f];

			pos = 2;
			buf = 0;
		}
	}
	if (pos != 2)
	{
		/*
		 * Leave if there is an overflow in the area allocated for the encoded
		 * string.
		 */
		if ((p - dst + 4) > dstlen)
			goto error;

		*p++ = _base64[(buf >> 18) & 0x3f];
		*p++ = _base64[(buf >> 12) & 0x3f];
		*p++ = (pos == 0) ? _base64[(buf >> 6) & 0x3f] : '=';
		*p++ = '=';
	}

	Assert((p - dst) <= dstlen);
	return p - dst;

error:
	memset(dst, 0, dstlen);
	return -1;
}

2.4 解码

设计妙处1：

• 使用b64lookup映射表直接反映射回原始下标值，而不用计算

static const int8 b64lookup[128] = {
	-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
	-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
	-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 62, -1, -1, -1, 63,
	52, 53, 54, 55, 56, 57, 58, 59, 60, 61, -1, -1, -1, -1, -1, -1,
	-1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
	15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, -1, -1, -1, -1, -1,
	-1, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
	41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, -1, -1, -1, -1, -1,
};

/*
 * pg_b64_decode
 *
 * Decode the given base64 string.  Returns the length of the decoded
 * string on success, and -1 in the event of an error with the result
 * buffer zeroed for safety.
 */
int
pg_b64_decode(const char *src, int len, char *dst, int dstlen)
{
	const char *srcend = src + len,
			   *s = src;
	char	   *p = dst;
	char		c;
	int			b = 0;
	uint32		buf = 0;
	int			pos = 0,
				end = 0;

	while (s < srcend)
	{
		c = *s++;

		/* Leave if a whitespace is found */
		if (c == ' ' || c == '\t' || c == '\n' || c == '\r')
			goto error;

		if (c == '=')
		{
			/* end sequence */
			if (!end)
			{
				if (pos == 2)
					end = 1;
				else if (pos == 3)
					end = 2;
				else
				{
					/*
					 * Unexpected "=" character found while decoding base64
					 * sequence.
					 */
					goto error;
				}
			}
			b = 0;
		}
		else
		{
			b = -1;
			if (c > 0 && c < 127)
				b = b64lookup[(unsigned char) c];
			if (b < 0)
			{
				/* invalid symbol found */
				goto error;
			}
		}
		/* add it to buffer */
		buf = (buf << 6) + b;
		pos++;
		if (pos == 4)
		{
			/*
			 * Leave if there is an overflow in the area allocated for the
			 * decoded string.
			 */
			if ((p - dst + 1) > dstlen)
				goto error;
			*p++ = (buf >> 16) & 255;

			if (end == 0 || end > 1)
			{
				/* overflow check */
				if ((p - dst + 1) > dstlen)
					goto error;
				*p++ = (buf >> 8) & 255;
			}
			if (end == 0 || end > 2)
			{
				/* overflow check */
				if ((p - dst + 1) > dstlen)
					goto error;
				*p++ = buf & 255;
			}
			buf = 0;
			pos = 0;
		}
	}

	if (pos != 0)
	{
		/*
		 * base64 end sequence is invalid.  Input data is missing padding, is
		 * truncated or is otherwise corrupted.
		 */
		goto error;
	}

	Assert((p - dst) <= dstlen);
	return p - dst;

error:
	memset(dst, 0, dstlen);
	return -1;
}