C++中的Base64解码片段

97

在C++中有免费的Base64解码代码片段可用吗?


另一个线程展示了如何使用boost进行base64编码/解码:https://dev59.com/dGkv5IYBdhLWcg3wiRao - Oded Itzhaky
1
使用Boost C++库进行Base64解码:http://stackoverflow.com/questions/34680998/attempt-to-decode-a-value-not-in-base64-char-set - ap6491
2
在Windows上,您可以使用atlenc.h或wincrypt.h来完成此操作。 - Hemaolle
13个回答

136

这是我对最初由René Nyffenegger编写的实现进行的修改。为什么我要修改呢?因为在我的看法中,使用std::string对象存储二进制数据似乎不太合适 ;)

base64.h:

#ifndef _BASE64_H_
#define _BASE64_H_

#include <vector>
#include <string>
typedef unsigned char BYTE;

std::string base64_encode(BYTE const* buf, unsigned int bufLen);
std::vector<BYTE> base64_decode(std::string const&);

#endif

base64.cpp:

#include "base64.h"
#include <iostream>

static const std::string base64_chars =
             "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
             "abcdefghijklmnopqrstuvwxyz"
             "0123456789+/";


static inline bool is_base64(BYTE c) {
  return (isalnum(c) || (c == '+') || (c == '/'));
}

std::string base64_encode(BYTE const* buf, unsigned int bufLen) {
  std::string ret;
  int i = 0;
  int j = 0;
  BYTE char_array_3[3];
  BYTE char_array_4[4];

  while (bufLen--) {
    char_array_3[i++] = *(buf++);
    if (i == 3) {
      char_array_4[0] = (char_array_3[0] & 0xfc) >> 2;
      char_array_4[1] = ((char_array_3[0] & 0x03) << 4) + ((char_array_3[1] & 0xf0) >> 4);
      char_array_4[2] = ((char_array_3[1] & 0x0f) << 2) + ((char_array_3[2] & 0xc0) >> 6);
      char_array_4[3] = char_array_3[2] & 0x3f;

      for(i = 0; (i <4) ; i++)
        ret += base64_chars[char_array_4[i]];
      i = 0;
    }
  }

  if (i)
  {
    for(j = i; j < 3; j++)
      char_array_3[j] = '\0';

    char_array_4[0] = (char_array_3[0] & 0xfc) >> 2;
    char_array_4[1] = ((char_array_3[0] & 0x03) << 4) + ((char_array_3[1] & 0xf0) >> 4);
    char_array_4[2] = ((char_array_3[1] & 0x0f) << 2) + ((char_array_3[2] & 0xc0) >> 6);
    char_array_4[3] = char_array_3[2] & 0x3f;

    for (j = 0; (j < i + 1); j++)
      ret += base64_chars[char_array_4[j]];

    while((i++ < 3))
      ret += '=';
  }

  return ret;
}

std::vector<BYTE> base64_decode(std::string const& encoded_string) {
  int in_len = encoded_string.size();
  int i = 0;
  int j = 0;
  int in_ = 0;
  BYTE char_array_4[4], char_array_3[3];
  std::vector<BYTE> ret;

  while (in_len-- && ( encoded_string[in_] != '=') && is_base64(encoded_string[in_])) {
    char_array_4[i++] = encoded_string[in_]; in_++;
    if (i ==4) {
      for (i = 0; i <4; i++)
        char_array_4[i] = base64_chars.find(char_array_4[i]);

      char_array_3[0] = (char_array_4[0] << 2) + ((char_array_4[1] & 0x30) >> 4);
      char_array_3[1] = ((char_array_4[1] & 0xf) << 4) + ((char_array_4[2] & 0x3c) >> 2);
      char_array_3[2] = ((char_array_4[2] & 0x3) << 6) + char_array_4[3];

      for (i = 0; (i < 3); i++)
          ret.push_back(char_array_3[i]);
      i = 0;
    }
  }

  if (i) {
    for (j = i; j <4; j++)
      char_array_4[j] = 0;

    for (j = 0; j <4; j++)
      char_array_4[j] = base64_chars.find(char_array_4[j]);

    char_array_3[0] = (char_array_4[0] << 2) + ((char_array_4[1] & 0x30) >> 4);
    char_array_3[1] = ((char_array_4[1] & 0xf) << 4) + ((char_array_4[2] & 0x3c) >> 2);
    char_array_3[2] = ((char_array_4[2] & 0x3) << 6) + char_array_4[3];

    for (j = 0; (j < i - 1); j++) ret.push_back(char_array_3[j]);
  }

  return ret;
}

以下是使用方法:

std::vector<BYTE> myData;
...
std::string encodedData = base64_encode(&myData[0], myData.size());
std::vector<BYTE> decodedData = base64_decode(encodedData);

2
谢谢。您的代码使用有任何限制吗?(我的个人情况是一个学术CFD代码)。 - Azrael3000
1
我试着使用你的方法解码一个jpg文件,通过将整个向量写入带有大小参数的CFile中,但并不奇怪,图像头部被损坏了,尽管大小是相等的。有更好的想法如何恢复图像文件吗? - masche
3
@masche:这是关于在字节级别上对任何类型的数据进行编码和解码。图像->原始数据(字节)->编码为base64字符串,然后再次转换为base64字符串->解码为原始数据(字节)->在其上构建一些输入流或对象或其他内容以再次将其作为图像处理... - LihO
1
抱歉 - 你的代码完美地运行了,甚至可以恢复图像。我太愚蠢了,只是把整个向量压缩到一个CFile中,这样是行不通的!如果我迭代向量并将每个单独的字节写入文件,它就可以工作了。也许文件流在这里是更好的解决方案。 - masche
8
在std::string中存储二进制数据有什么问题? - GaspardP
显示剩余6条评论

107

请参考使用C++编码和解码base64

这是该页面的实现代码:

/*
   base64.cpp and base64.h

   Copyright (C) 2004-2008 René Nyffenegger

   This source code is provided 'as-is', without any express or implied
   warranty. In no event will the author be held liable for any damages
   arising from the use of this software.

   Permission is granted to anyone to use this software for any purpose,
   including commercial applications, and to alter it and redistribute it
   freely, subject to the following restrictions:

   1. The origin of this source code must not be misrepresented; you must not
      claim that you wrote the original source code. If you use this source code
      in a product, an acknowledgment in the product documentation would be
      appreciated but is not required.

   2. Altered source versions must be plainly marked as such, and must not be
      misrepresented as being the original source code.

   3. This notice may not be removed or altered from any source distribution.

   René Nyffenegger rene.nyffenegger@adp-gmbh.ch

*/

static const std::string base64_chars =
             "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
             "abcdefghijklmnopqrstuvwxyz"
             "0123456789+/";


static inline bool is_base64(unsigned char c) {
  return (isalnum(c) || (c == '+') || (c == '/'));
}

std::string base64_encode(unsigned char const* bytes_to_encode, unsigned int in_len) {
  std::string ret;
  int i = 0;
  int j = 0;
  unsigned char char_array_3[3];
  unsigned char char_array_4[4];

  while (in_len--) {
    char_array_3[i++] = *(bytes_to_encode++);
    if (i == 3) {
      char_array_4[0] = (char_array_3[0] & 0xfc) >> 2;
      char_array_4[1] = ((char_array_3[0] & 0x03) << 4) + ((char_array_3[1] & 0xf0) >> 4);
      char_array_4[2] = ((char_array_3[1] & 0x0f) << 2) + ((char_array_3[2] & 0xc0) >> 6);
      char_array_4[3] = char_array_3[2] & 0x3f;

      for(i = 0; (i <4) ; i++)
        ret += base64_chars[char_array_4[i]];
      i = 0;
    }
  }

  if (i)
  {
    for(j = i; j < 3; j++)
      char_array_3[j] = '\0';

    char_array_4[0] = (char_array_3[0] & 0xfc) >> 2;
    char_array_4[1] = ((char_array_3[0] & 0x03) << 4) + ((char_array_3[1] & 0xf0) >> 4);
    char_array_4[2] = ((char_array_3[1] & 0x0f) << 2) + ((char_array_3[2] & 0xc0) >> 6);
    char_array_4[3] = char_array_3[2] & 0x3f;

    for (j = 0; (j < i + 1); j++)
      ret += base64_chars[char_array_4[j]];

    while((i++ < 3))
      ret += '=';

  }

  return ret;

}
std::string base64_decode(std::string const& encoded_string) {
  int in_len = encoded_string.size();
  int i = 0;
  int j = 0;
  int in_ = 0;
  unsigned char char_array_4[4], char_array_3[3];
  std::string ret;

  while (in_len-- && ( encoded_string[in_] != '=') && is_base64(encoded_string[in_])) {
    char_array_4[i++] = encoded_string[in_]; in_++;
    if (i ==4) {
      for (i = 0; i <4; i++)
        char_array_4[i] = base64_chars.find(char_array_4[i]);

      char_array_3[0] = (char_array_4[0] << 2) + ((char_array_4[1] & 0x30) >> 4);
      char_array_3[1] = ((char_array_4[1] & 0xf) << 4) + ((char_array_4[2] & 0x3c) >> 2);
      char_array_3[2] = ((char_array_4[2] & 0x3) << 6) + char_array_4[3];

      for (i = 0; (i < 3); i++)
        ret += char_array_3[i];
      i = 0;
    }
  }

  if (i) {
    for (j = i; j <4; j++)
      char_array_4[j] = 0;

    for (j = 0; j <4; j++)
      char_array_4[j] = base64_chars.find(char_array_4[j]);

    char_array_3[0] = (char_array_4[0] << 2) + ((char_array_4[1] & 0x30) >> 4);
    char_array_3[1] = ((char_array_4[1] & 0xf) << 4) + ((char_array_4[2] & 0x3c) >> 2);
    char_array_3[2] = ((char_array_4[2] & 0x3) << 6) + char_array_4[3];

    for (j = 0; (j < i - 1); j++) ret += char_array_3[j];
  }

  return ret;
}

5
既然我们已经知道了in_len,也就知道了ret的长度,为什么不在初始化时就给它一个固定的长度呢?这样我们就可以避免所有这些不必要的字符串拼接。 - Ray
在解码函数中,if(i)后面的第一个for循环是不必要的。如果char_array_4中的字节填充为0,则find()返回-1,但根本没有使用它。因此,第二个for循环可以是:for(j = 0; j < i; j++)。另外,带有char_array_3[2]的第三行是完全无用的,可以省略。为什么?因为这个第二段只处理剩余的(当i不为零时),而原始文本中只能是一个或两个字节(如果是三个字节,则不会有剩余,因为三个字节可以顺利地编码为4个字符)。 - Ingmar
7
谷歌并不总是你的朋友。这个实现几乎是你可能选择的最糟糕的一个。参见此链接:https://dev59.com/IXRC5IYBdhLWcg3wUfJ2#41094722 - DaedalusAlpha
这段代码有漏洞。将解码后的值存储在字符串中是错误的!想象一下,当你编码一个字节数组时,某些值可能是'\0',编码部分完全没有问题。但是当你反转回来时,该值将被截断为'\0'。请参考@LihO的答案。 - Hainan Zhao
@HainanZhao C++字符串(与C字符串相比)可以很好地处理NUL字节。因此,除非使用结果字符串的c_str()方法,否则一切都很好。 - Kai Petzke

63

这里有几个代码片段。然而,这个代码片段既紧凑又高效,且符合C++11标准:

static std::string base64_encode(const std::string &in) {

    std::string out;

    int val = 0, valb = -6;
    for (uchar c : in) {
        val = (val << 8) + c;
        valb += 8;
        while (valb >= 0) {
            out.push_back("ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"[(val>>valb)&0x3F]);
            valb -= 6;
        }
    }
    if (valb>-6) out.push_back("ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"[((val<<8)>>(valb+8))&0x3F]);
    while (out.size()%4) out.push_back('=');
    return out;
}

static std::string base64_decode(const std::string &in) {

    std::string out;

    std::vector<int> T(256,-1);
    for (int i=0; i<64; i++) T["ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"[i]] = i;

    int val=0, valb=-8;
    for (uchar c : in) {
        if (T[c] == -1) break;
        val = (val << 6) + T[c];
        valb += 6;
        if (valb >= 0) {
            out.push_back(char((val>>valb)&0xFF));
            valb -= 8;
        }
    }
    return out;
}

1
我应该收回那句话...我不应该在调试模式下进行性能测试,抱歉。但至少比已接受的解决方案快。 - Marco Freudenberger
6
将“int val”移出其范围的位移是未定义行为。使用“unsigned val = 0;int valb = ...”是正确的。 - Kevin Yin
2
typedef unsigned char uchar; - Emrexdy
1
out.reserve(8 * (1 + in.size() / 6)); 输出.保留(8 * (1 + in.size() / 6)); - Larytet

41

我认为这一个更好:

#include <string>

static const char* B64chars = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";

static const int B64index[256] =
{
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  62, 63, 62, 62, 63,
    52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 0,  0,  0,  0,  0,  0,
    0,  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, 0,  0,  0,  0,  63,
    0,  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
};

const std::string b64encode(const void* data, const size_t &len)
{
    std::string result((len + 2) / 3 * 4, '=');
    unsigned char *p = (unsigned  char*) data;
    char *str = &result[0];
    size_t j = 0, pad = len % 3;
    const size_t last = len - pad;

    for (size_t i = 0; i < last; i += 3)
    {
        int n = int(p[i]) << 16 | int(p[i + 1]) << 8 | p[i + 2];
        str[j++] = B64chars[n >> 18];
        str[j++] = B64chars[n >> 12 & 0x3F];
        str[j++] = B64chars[n >> 6 & 0x3F];
        str[j++] = B64chars[n & 0x3F];
    }
    if (pad)  /// Set padding
    {
        int n = --pad ? int(p[last]) << 8 | p[last + 1] : p[last];
        str[j++] = B64chars[pad ? n >> 10 & 0x3F : n >> 2];
        str[j++] = B64chars[pad ? n >> 4 & 0x03F : n << 4 & 0x3F];
        str[j++] = pad ? B64chars[n << 2 & 0x3F] : '=';
    }
    return result;
}

const std::string b64decode(const void* data, const size_t &len)
{
    if (len == 0) return "";

    unsigned char *p = (unsigned char*) data;
    size_t j = 0,
        pad1 = len % 4 || p[len - 1] == '=',
        pad2 = pad1 && (len % 4 > 2 || p[len - 2] != '=');
    const size_t last = (len - pad1) / 4 << 2;
    std::string result(last / 4 * 3 + pad1 + pad2, '\0');
    unsigned char *str = (unsigned char*) &result[0];

    for (size_t i = 0; i < last; i += 4)
    {
        int n = B64index[p[i]] << 18 | B64index[p[i + 1]] << 12 | B64index[p[i + 2]] << 6 | B64index[p[i + 3]];
        str[j++] = n >> 16;
        str[j++] = n >> 8 & 0xFF;
        str[j++] = n & 0xFF;
    }
    if (pad1)
    {
        int n = B64index[p[last]] << 18 | B64index[p[last + 1]] << 12;
        str[j++] = n >> 16;
        if (pad2)
        {
            n |= B64index[p[last + 2]] << 6;
            str[j++] = n >> 8 & 0xFF;
        }
    }
    return result;
}

std::string b64encode(const std::string& str)
{
    return b64encode(str.c_str(), str.size());
}

std::string b64decode(const std::string& str64)
{
    return b64decode(str64.c_str(), str64.size());
}

由于Jens Alfke指出了性能问题,我对这篇文章进行了一些修改。现在它比以前快多了。另一个优点是能够平稳地处理损坏的数据。

最终版本:虽然在这类问题中,速度似乎是过度追求,但为了好玩,我进行了其他一些修改,使其成为目前我所知道的最快算法。特别感谢GaspardP的宝贵建议和很好的基准测试。


1
所有对 strchr 的调用都会减慢解码器的速度 - 每个字节解码平均循环32次。大多数解决方案使用256项查找表来避免这种情况,这样会快得多。 - Jens Alfke
4
我曾在https://dev59.com/IXRC5IYBdhLWcg3wUfJ2#41094722上发表评论——尽管这段代码不是最快的编码方式,但它是最快的解码方式(与其他16个实现进行了比较)。 - GaspardP
4
在解码器中,你可以通过使用char*代替std::string来进一步提高性能(在我的测试中可以提高10-15%)。只需执行char * out = &str[0],然后使用out[j ++]代替str[j ++]即可。这样做可以避免由于std::string :: operator []执行的不必要检查。另外,避免使用最后一个push_back,因为它可能非常昂贵,可以先分配多一个字节(std::string str; str.resize(3*((len+3)/4));),然后在每个地方都使用out[j ++],最后使用str.resize(j); - GaspardP
2
你在上次的编辑中添加了一个内存泄漏,更不用说缓冲区复制了。不要在没有使用delete的情况下使用new。实际上,根本不要使用new。@Gaspard,我不知道std::string::operator[]是否会进行任何“不必要的检查”(事实上,我相当确定至少在发布版本中没有),但如果真的绝望,你可以使用vector<char> - 无论如何,你的替换代码并不是Gaspard建议的,他仍然使用字符串来完成所有操作,并且无论如何都是安全/快速的,尽管我认为这是不必要的 ;) - Lightness Races in Orbit
8
谢谢你的留言。请在编码器中使用 unsigned char *p,而不是 char *p。如果我的输入包含字节 >= 0x80,则我会得到损坏的base64字符串。添加 unsigned 后,似乎一切都正常了。 - Denis Golovkin
显示剩余4条评论

18

使用base-n 迷你库,您可以执行以下操作:

some_data_t in[] { ... };
constexpr int len = sizeof(in)/sizeof(in[0]);

std::string encoded;
bn::encode_b64(in, in + len, std::back_inserter(encoded));

some_data_t out[len];
bn::decode_b64(encoded.begin(), encoded.end(), out);

这个API是通用的,基于迭代器的。

声明:我是作者。


3
std::sizesizeof 更好。 - Lightness Races in Orbit
1
确实,从c++17开始,这显然是正确的做法。 - azawadzki
1
自从一开始就是这样做的正确方式。只是在 C++17 之前,你必须自己实现它(但实际上只需要一行代码)。在 C++ 中,sizeof hack 永远都不可以。 - Lightness Races in Orbit

16
根据GaspardP所做的优秀比较,我不会选择这个解决方案。它不是最糟糕的,但也不是最好的。唯一有点的就是它可能更容易理解。
我发现其他两个答案很难理解。它们还在我的编译器中产生了一些警告,并且在解码部分使用find函数应该会导致相当糟糕的效率。因此我决定自己动手写。
头文件:
#ifndef _BASE64_H_
#define _BASE64_H_

#include <vector>
#include <string>
typedef unsigned char BYTE;

class Base64
{
public:
    static std::string encode(const std::vector<BYTE>& buf);
    static std::string encode(const BYTE* buf, unsigned int bufLen);
    static std::vector<BYTE> decode(std::string encoded_string);
};

#endif

正文:

static const BYTE from_base64[] = {    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
                                    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
                                    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,  62, 255,  62, 255,  63,
                                     52,  53,  54,  55,  56,  57,  58,  59,  60,  61, 255, 255, 255, 255, 255, 255,
                                    255,   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, 255, 255, 255, 255,  63,
                                    255,  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, 255, 255, 255, 255, 255};

static const char to_base64[] =
             "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
             "abcdefghijklmnopqrstuvwxyz"
             "0123456789+/";


std::string Base64::encode(const std::vector<BYTE>& buf)
{
    if (buf.empty())
        return ""; // Avoid dereferencing buf if it's empty
    return encode(&buf[0], (unsigned int)buf.size());
}

std::string Base64::encode(const BYTE* buf, unsigned int bufLen)
{
    // Calculate how many bytes that needs to be added to get a multiple of 3
    size_t missing = 0;
    size_t ret_size = bufLen;
    while ((ret_size % 3) != 0)
    {
        ++ret_size;
        ++missing;
    }

    // Expand the return string size to a multiple of 4
    ret_size = 4*ret_size/3;

    std::string ret;
    ret.reserve(ret_size);

    for (unsigned int i=0; i<ret_size/4; ++i)
    {
        // Read a group of three bytes (avoid buffer overrun by replacing with 0)
        size_t index = i*3;
        BYTE b3[3];
        b3[0] = (index+0 < bufLen) ? buf[index+0] : 0;
        b3[1] = (index+1 < bufLen) ? buf[index+1] : 0;
        b3[2] = (index+2 < bufLen) ? buf[index+2] : 0;

        // Transform into four base 64 characters
        BYTE b4[4];
        b4[0] =                            ((b3[0] & 0xfc) >> 2);
        b4[1] = ((b3[0] & 0x03) << 4) +    ((b3[1] & 0xf0) >> 4);
        b4[2] = ((b3[1] & 0x0f) << 2) +    ((b3[2] & 0xc0) >> 6);
        b4[3] = ((b3[2] & 0x3f) << 0);

        // Add the base 64 characters to the return value
        ret.push_back(to_base64[b4[0]]);
        ret.push_back(to_base64[b4[1]]);
        ret.push_back(to_base64[b4[2]]);
        ret.push_back(to_base64[b4[3]]);
    }

    // Replace data that is invalid (always as many as there are missing bytes)
    for (size_t i=0; i<missing; ++i)
        ret[ret_size - i - 1] = '=';

    return ret;
}

std::vector<BYTE> Base64::decode(std::string encoded_string)
{
    // Make sure string length is a multiple of 4
    while ((encoded_string.size() % 4) != 0)
        encoded_string.push_back('=');

    size_t encoded_size = encoded_string.size();
    std::vector<BYTE> ret;
    ret.reserve(3*encoded_size/4);

    for (size_t i=0; i<encoded_size; i += 4)
    {
        // Get values for each group of four base 64 characters
        BYTE b4[4];
        b4[0] = (encoded_string[i+0] <= 'z') ? from_base64[encoded_string[i+0]] : 0xff;
        b4[1] = (encoded_string[i+1] <= 'z') ? from_base64[encoded_string[i+1]] : 0xff;
        b4[2] = (encoded_string[i+2] <= 'z') ? from_base64[encoded_string[i+2]] : 0xff;
        b4[3] = (encoded_string[i+3] <= 'z') ? from_base64[encoded_string[i+3]] : 0xff;

        // Transform into a group of three bytes
        BYTE b3[3];
        b3[0] = ((b4[0] & 0x3f) << 2) + ((b4[1] & 0x30) >> 4);
        b3[1] = ((b4[1] & 0x0f) << 4) + ((b4[2] & 0x3c) >> 2);
        b3[2] = ((b4[2] & 0x03) << 6) + ((b4[3] & 0x3f) >> 0);

        // Add the byte to the return value if it isn't part of an '=' character (indicated by 0xff)
        if (b4[1] != 0xff) ret.push_back(b3[0]);
        if (b4[2] != 0xff) ret.push_back(b3[1]);
        if (b4[3] != 0xff) ret.push_back(b3[2]);
    }

    return ret;
}

使用方法:

BYTE buf[] = "ABCD";
std::string encoded = Base64::encode(buf, 4);
// encoded = "QUJDRA=="
std::vector<BYTE> decoded = Base64::decode(encoded);

这里的一个额外优势是,decode函数还可以解码Base64编码的URL变体。


2
加分项是不使用find()函数,并为输出使用reserve()。因为您将输入作为副本,所以少得了一点分(如果需要,可以在末尾添加=)。如果它是无复制的东西,那就更好了。 - elegant dice
3
我最喜欢这个回答了……对于行数来说也非常好。如果没有使用find()函数,并且对输出进行反转(),那么可以得到额外的加分。 我想改进的地方(这些将增加代码量): 您将输入作为副本处理(因此如果需要,可以在末尾添加=)。如果它是一个无复制的内容,那就更好了。 也可以将其编写为普通的函数--不需要类。 在引用buf[0]之前应该检查空缓冲向量。 并添加一个接口来将数据写出到引用中(这样调用者可以重用内存)。 - elegant dice
1
感谢您的反馈,我采用按值传递字符串的原因是,如果我能始终保证它具有有效长度,那么它可以简化代码。我认为在大多数情况下,RVO应该会防止字符串复制,所以这不应该是一个问题。至于解引用buf [0] - 很好的发现,我会修复它 :) - DaedalusAlpha
1
我完成了自己修改的版本,为了让你满意,我会在这里发布答案...我通过添加一个测试来消除作为副本的字符串的需要(你已经进行测试,所以它并不会增加任何性能问题)。 - elegant dice
1
我可能错了,但是从from_base64[]数组来看,索引60和61的值不应该交换吗? 我猜想的想法是通过返回255忽略"<"字符(索引=60),并通过返回0识别填充字符"="(索引=61)。 - pbyhistorian
显示剩余2条评论

10

一个使用更紧凑的查找表和C++17特性的小变化:

std::string base64_decode(const std::string_view in) {
  // table from '+' to 'z'
  const uint8_t lookup[] = {
      62,  255, 62,  255, 63,  52,  53, 54, 55, 56, 57, 58, 59, 60, 61, 255,
      255, 0,   255, 255, 255, 255, 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,
      255, 255, 255, 255, 63,  255, 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};
  static_assert(sizeof(lookup) == 'z' - '+' + 1);

  std::string out;
  int val = 0, valb = -8;
  for (uint8_t c : in) {
    if (c < '+' || c > 'z')
      break;
    c -= '+';
    if (lookup[c] >= 64)
      break;
    val = (val << 6) + lookup[c];
    valb += 6;
    if (valb >= 0) {
      out.push_back(char((val >> valb) & 0xFF));
      valb -= 8;
    }
  }
  return out;
}

如果您没有std::string_view,请尝试使用std::experimental::string_view替代。


5

我对 DaedalusAlpha 的答案进行了改进:

它通过一些测试来避免复制参数。

使用 uint8_t 替代 BYTE。

添加了一些方便处理字符串的函数,尽管通常输入数据是二进制的并且可能包含零字节,因此通常不应将其作为字符串操作(这通常意味着空终止数据)。

还添加了一些强制转换以解决编译器警告(至少在 GCC 上,我还没有在 MSVC 上运行它)。

文件的一部分是 base64.hpp

void base64_encode(string & out, const vector<uint8_t>& buf);
void base64_encode(string & out, const uint8_t* buf, size_t bufLen);
void base64_encode(string & out, string const& buf);

void base64_decode(vector<uint8_t> & out, string const& encoded_string);

// Use this if you know the output should be a valid string
void base64_decode(string & out, string const& encoded_string);

文件 base64.cpp:

static const uint8_t from_base64[128] = {
    // 8 rows of 16 = 128
    // Note: only requires 123 entries, as we only lookup for <= z , which z=122

    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,  62, 255,  62, 255,  63,
     52,  53,  54,  55,  56,  57,  58,  59,  60,  61, 255, 255,   0, 255, 255, 255,
    255,   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, 255, 255, 255, 255,  63,
    255,  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, 255, 255, 255, 255, 255
};

static const char to_base64[65] =
    "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
    "abcdefghijklmnopqrstuvwxyz"
    "0123456789+/";


void base64_encode(string & out, string const& buf)
{
   if (buf.empty())
      base64_encode(out, NULL, 0);
   else
      base64_encode(out, reinterpret_cast<uint8_t const*>(&buf[0]), buf.size());
}


void base64_encode(string & out, std::vector<uint8_t> const& buf)
{
   if (buf.empty())
      base64_encode(out, NULL, 0);
   else
      base64_encode(out, &buf[0], buf.size());
}

void base64_encode(string & ret, uint8_t const* buf, size_t bufLen)
{
   // Calculate how many bytes that needs to be added to get a multiple of 3
   size_t missing = 0;
   size_t ret_size = bufLen;
   while ((ret_size % 3) != 0)
   {
      ++ret_size;
      ++missing;
   }

   // Expand the return string size to a multiple of 4
   ret_size = 4*ret_size/3;

   ret.clear();
   ret.reserve(ret_size);

   for (size_t i = 0; i < ret_size/4; ++i)
   {
      // Read a group of three bytes (avoid buffer overrun by replacing with 0)
      const size_t index = i*3;
      const uint8_t b3_0 = (index+0 < bufLen) ? buf[index+0] : 0;
      const uint8_t b3_1 = (index+1 < bufLen) ? buf[index+1] : 0;
      const uint8_t b3_2 = (index+2 < bufLen) ? buf[index+2] : 0;

      // Transform into four base 64 characters
      const uint8_t b4_0 =                        ((b3_0 & 0xfc) >> 2);
      const uint8_t b4_1 = ((b3_0 & 0x03) << 4) + ((b3_1 & 0xf0) >> 4);
      const uint8_t b4_2 = ((b3_1 & 0x0f) << 2) + ((b3_2 & 0xc0) >> 6);
      const uint8_t b4_3 = ((b3_2 & 0x3f) << 0);

      // Add the base 64 characters to the return value
      ret.push_back(to_base64[b4_0]);
      ret.push_back(to_base64[b4_1]);
      ret.push_back(to_base64[b4_2]);
      ret.push_back(to_base64[b4_3]);
   }

   // Replace data that is invalid (always as many as there are missing bytes)
   for (size_t i = 0; i != missing; ++i)
      ret[ret_size - i - 1] = '=';
}


template <class Out>
void base64_decode_any( Out & ret, std::string const& in)
{
   typedef typename Out::value_type T;

   // Make sure the *intended* string length is a multiple of 4
   size_t encoded_size = in.size();

   while ((encoded_size % 4) != 0)
      ++encoded_size;

   const size_t N = in.size();
   ret.clear();
   ret.reserve(3*encoded_size/4);

   for (size_t i = 0; i < encoded_size; i += 4)
   {
      // Note: 'z' == 122

      // Get values for each group of four base 64 characters
      const uint8_t b4_0 = (            in[i+0] <= 'z') ? from_base64[static_cast<uint8_t>(in[i+0])] : 0xff;
      const uint8_t b4_1 = (i+1 < N and in[i+1] <= 'z') ? from_base64[static_cast<uint8_t>(in[i+1])] : 0xff;
      const uint8_t b4_2 = (i+2 < N and in[i+2] <= 'z') ? from_base64[static_cast<uint8_t>(in[i+2])] : 0xff;
      const uint8_t b4_3 = (i+3 < N and in[i+3] <= 'z') ? from_base64[static_cast<uint8_t>(in[i+3])] : 0xff;

      // Transform into a group of three bytes
      const uint8_t b3_0 = ((b4_0 & 0x3f) << 2) + ((b4_1 & 0x30) >> 4);
      const uint8_t b3_1 = ((b4_1 & 0x0f) << 4) + ((b4_2 & 0x3c) >> 2);
      const uint8_t b3_2 = ((b4_2 & 0x03) << 6) + ((b4_3 & 0x3f) >> 0);

      // Add the byte to the return value if it isn't part of an '=' character (indicated by 0xff)
      if (b4_1 != 0xff) ret.push_back( static_cast<T>(b3_0) );
      if (b4_2 != 0xff) ret.push_back( static_cast<T>(b3_1) );
      if (b4_3 != 0xff) ret.push_back( static_cast<T>(b3_2) );
   }
}

void base64_decode(vector<uint8_t> & out, string const& encoded_string)
{
   base64_decode_any(out, encoded_string);
}

void base64_decode(string & out, string const& encoded_string)
{
   base64_decode_any(out, encoded_string);
}

1
我之前不知道在C++中,&&和||有一个定义好的评估顺序,所以今天我学到了新东西。在像这样的情况下,当你想要检查索引的条件,但同时又确保索引不超出范围时,它非常有用。 - DaedalusAlpha
1
是的,我一直使用那种技巧。 - elegant dice

3

我的版本是一个简单快速的 C++Builder Base64 编码器(解码器)。

// ---------------------------------------------------------------------------
UnicodeString __fastcall TExample::Base64Encode(void *data, int length)
{
    if (length <= 0)
        return L"";
    static const char set[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
    unsigned char *in = (unsigned char*)data;
    char *pos, *out = pos = new char[((length - 1) / 3 + 1) << 2];
    while ((length -= 3) >= 0)
    {
        pos[0] = set[in[0] >> 2];
        pos[1] = set[((in[0] & 0x03) << 4) | (in[1] >> 4)];
        pos[2] = set[((in[1] & 0x0F) << 2) | (in[2] >> 6)];
        pos[3] = set[in[2] & 0x3F];
        pos += 4;
        in += 3;
    };
    if ((length & 2) != 0)
    {
        pos[0] = set[in[0] >> 2];
        if ((length & 1) != 0)
        {
            pos[1] = set[((in[0] & 0x03) << 4) | (in[1] >> 4)];
            pos[2] = set[(in[1] & 0x0F) << 2];
        }
        else
        {
            pos[1] = set[(in[0] & 0x03) << 4];
            pos[2] = '=';
        };
        pos[3] = '=';
        pos += 4;
    };
    UnicodeString code = UnicodeString(out, pos - out);
    delete[] out;
    return code;
};

// ---------------------------------------------------------------------------
int __fastcall TExample::Base64Decode(const UnicodeString &code, unsigned char **data)
{
    int length;
    if (((length = code.Length()) == 0) || ((length & 3) != 0))
        return 0;
    wchar_t *str = code.c_str();
    unsigned char *pos, *out = pos = new unsigned char[(length >> 2) * 3];
    while (*str != 0)
    {
        length = -1;
        int shift = 18, bits = 0;
        do
        {
            wchar_t s = str[++length];
            if ((s >= L'A') && (s <= L'Z'))
                bits |= (s - L'A') << shift;
            else if ((s >= L'a') && (s <= L'z'))
                   bits |= (s - (L'a' - 26)) << shift;
            else if (((s >= L'0') && (s <= L'9')))
                   bits |= (s - (L'0' - 52)) << shift;
            else if (s == L'+')
                   bits |= 62 << shift;
            else if (s == L'/')
                   bits |= 63 << shift;
            else if (s == L'=')
            {
                length--;
                break;
            }
            else
            {
                delete[] out;
                return 0;
            };
        }
        while ((shift -= 6) >= 0);
        pos[0] = bits >> 16;
        pos[1] = bits >> 8;
        pos[2] = bits;
        pos += length;
        str += 4;
    };
    *data = out;
    return pos - out;
};
//---------------------------------------------------------------------------

2
我使用这个:
class BinaryVector {
public:
    std::vector<char> bytes;

    uint64_t bit_count = 0;

public:
    /* Add a bit to the end */
    void push_back(bool bit);

    /* Return false if character is unrecognized */
    bool pushBase64Char(char b64_c);
};

void BinaryVector::push_back(bool bit)
{
    if (!bit_count || bit_count % 8 == 0) {
        bytes.push_back(bit << 7);
    }
    else {
        uint8_t next_bit = 8 - (bit_count % 8) - 1;
        bytes[bit_count / 8] |= bit << next_bit;
    }
    bit_count++;
}

/* Converts one Base64 character to 6 bits */
bool BinaryVector::pushBase64Char(char c)
{
    uint8_t d;

    // A to Z
    if (c > 0x40 && c < 0x5b) {
        d = c - 65;  // Base64 A is 0
    }
    // a to z
    else if (c > 0x60 && c < 0x7b) {
        d = c - 97 + 26;  // Base64 a is 26
    }
    // 0 to 9
    else if (c > 0x2F && c < 0x3a) {
        d = c - 48 + 52;  // Base64 0 is 52
    }
    else if (c == '+') {
        d = 0b111110;
    }
    else if (c == '/') {
        d = 0b111111;
    }
    else if (c == '=') {
        d = 0;
    }
    else {
        return false;
    }

    push_back(d & 0b100000);
    push_back(d & 0b010000);
    push_back(d & 0b001000);
    push_back(d & 0b000100);
    push_back(d & 0b000010);
    push_back(d & 0b000001);

    return true;
}

bool loadBase64(std::vector<char>& b64_bin, BinaryVector& vec)
{
    for (char& c : b64_bin) {
        if (!vec.pushBase64Char(c)) {
            return false;
        }
    }
    return true;
}


使用vec.bytes来访问转换后的数据。

网页内容由stack overflow 提供, 点击上面的
可以查看英文原文,
原文链接