Base64 encoding: turning binary data into ASCII text for the purpose of saving it to text files like XML, or transmitting it over protocols like HTTP, or embedding it into web page files, and many other purposed. That’s the basic idea behind it. For every 3 bytes of input you get 4 bytes of output, so it’s not crazy inflated.
I was looking for a library that has built in, easy to use and clean base64 encoding and decoding functions but didn’t really find anything to my liking. So I looked for a reference implementation and found one at wikibooks.org. Their C++ implementation (released to public domain so I could freely use and modify it) was my starting point. I beautified the code and brought it closer to modern C++ 🙂 So now you have a header only, clean base64 encode and decode functions you can use in your projects: base64.hpp.
I wrote a program that encodes and decodes an input string, checks the original against the decoded one, and also checks the encoded base64 text against reference base64 string taken from wiki. The implementation checks out and produces correct encoded strings and decoded data 🙂 Below is the test program.
|
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 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 52 53 54 55 |
#include <iostream> #include <string> #include <vector> #include <iterator> #include "base64.hpp" #include "ascii_escape_code.hpp" using namespace std; using namespace ascii_escape_code; int main(int argc, char** argv) { string input { "Man is distinguished, not only by his reason, but by this singular passion from other animals, " "which is a lust of the mind, that by a perseverance of delight in the continued and indefatigable " "generation of knowledge, exceeds the short vehemence of any carnal pleasure." }; string reference { "TWFuIGlzIGRpc3Rpbmd1aXNoZWQsIG5vdCBvbmx5IGJ5IGhpcyByZWFzb24sIGJ1dCBieSB0aGlz" "IHNpbmd1bGFyIHBhc3Npb24gZnJvbSBvdGhlciBhbmltYWxzLCB3aGljaCBpcyBhIGx1c3Qgb2Yg" "dGhlIG1pbmQsIHRoYXQgYnkgYSBwZXJzZXZlcmFuY2Ugb2YgZGVsaWdodCBpbiB0aGUgY29udGlu" "dWVkIGFuZCBpbmRlZmF0aWdhYmxlIGdlbmVyYXRpb24gb2Yga25vd2xlZGdlLCBleGNlZWRzIHRo" "ZSBzaG9ydCB2ZWhlbWVuY2Ugb2YgYW55IGNhcm5hbCBwbGVhc3VyZS4=" }; try { vector<base64::byte> data(begin(input), end(input)); cout << bold << bright_red << "Input: " << reset << italic << input << reset << endl << endl; cout << bold << bright_red << "Reference: " << reset << reference << endl << endl; auto encoded = base64::encode(data); cout << bold << bright_red << "Encoded: " << reset << encoded << endl << endl; if(encoded != reference) throw runtime_error("Oh snap! Encoded data does not match reference!"); else cout << bright_red << "Encoded data matches reference :o)" << reset << endl << endl; auto decoded = base64::decode(encoded); string s2(begin(decoded), end(decoded)); cout << bold << bright_red << "Decoded: " << reset << italic << s2 << reset << endl << endl; if(data != decoded) throw runtime_error("Oh snap! Input data does not match decoded!"); else cout << bright_red << "Decoded data matches original :o)" << reset << endl << endl; } catch(exception& e) { cerr << slow_blink << bright_red << e.what() << reset << endl << endl; } return 1; } |
Input: Man is distinguished, not only by his reason, but by this singular passion from other animals, which is a lust of the mind, that by a perseverance of delight in the continued and indefatigable generation of knowledge, exceeds the short vehemence of any carnal pleasure.
Reference:TWFuIGlzIGRpc3Rpbmd1aXNoZWQsIG5vdCBvbmx5IGJ5IGhpcyByZWFzb24sIGJ 1dCBieSB0aGlzIHNpbmd1bGFyIHBhc3Npb24gZnJvbSBvdGhlciBhbmltYWxzLCB3 aGljaCBpcyBhIGx1c3Qgb2YgdGhlIG1pbmQsIHRoYXQgYnkgYSBwZXJzZXZlcmF uY2Ugb2YgZGVsaWdodCBpbiB0aGUgY29udGludWVkIGFuZCBpbmRlZmF0aWd hYmxlIGdlbmVyYXRpb24gb2Yga25vd2xlZGdlLCBleGNlZWRzIHRoZSBzaG9ydC B2ZWhlbWVuY2Ugb2YgYW55IGNhcm5hbCBwbGVhc3VyZS4=
Encoded:TWFuIGlzIGRpc3Rpbmd1aXNoZWQsIG5vdCBvbmx5IGJ5IGhpcyByZWFzb24sIGJ 1dCBieSB0aGlzIHNpbmd1bGFyIHBhc3Npb24gZnJvbSBvdGhlciBhbmltYWxzLCB3 aGljaCBpcyBhIGx1c3Qgb2YgdGhlIG1pbmQsIHRoYXQgYnkgYSBwZXJzZXZlcmF uY2Ugb2YgZGVsaWdodCBpbiB0aGUgY29udGludWVkIGFuZCBpbmRlZmF0aWd hYmxlIGdlbmVyYXRpb24gb2Yga25vd2xlZGdlLCBleGNlZWRzIHRoZSBzaG9ydC B2ZWhlbWVuY2Ugb2YgYW55IGNhcm5hbCBwbGVhc3VyZS4=
Program output.
Encoded data matches reference :o)
Decoded: Man is distinguished, not only by his reason, but by this singular passion from other animals, which is a lust of the mind, that by a perseverance of delight in the continued and indefatigable generation of knowledge, exceeds the short vehemence of any carnal pleasure.
Decoded data matches original :o)
And here is the encoder and decoder function implementation.
|
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 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 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 |
#pragma once #include <string> #include <vector> #include <stdexcept> #include <cstdint> namespace base64 { inline static const char kEncodeLookup[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; inline static const char kPadCharacter = '='; using byte = std::uint8_t; inline std::string encode(const std::vector<byte>& input) { std::string encoded; encoded.reserve(((input.size() / 3) + (input.size() % 3 > 0)) * 4); std::uint32_t temp{}; auto it = input.begin(); for(std::size_t i = 0; i < input.size() / 3; ++i) { temp = (*it++) << 16; temp += (*it++) << 8; temp += (*it++); encoded.append(1, kEncodeLookup[(temp & 0x00FC0000) >> 18]); encoded.append(1, kEncodeLookup[(temp & 0x0003F000) >> 12]); encoded.append(1, kEncodeLookup[(temp & 0x00000FC0) >> 6 ]); encoded.append(1, kEncodeLookup[(temp & 0x0000003F) ]); } switch(input.size() % 3) { case 1: temp = (*it++) << 16; encoded.append(1, kEncodeLookup[(temp & 0x00FC0000) >> 18]); encoded.append(1, kEncodeLookup[(temp & 0x0003F000) >> 12]); encoded.append(2, kPadCharacter); break; case 2: temp = (*it++) << 16; temp += (*it++) << 8; encoded.append(1, kEncodeLookup[(temp & 0x00FC0000) >> 18]); encoded.append(1, kEncodeLookup[(temp & 0x0003F000) >> 12]); encoded.append(1, kEncodeLookup[(temp & 0x00000FC0) >> 6 ]); encoded.append(1, kPadCharacter); break; } return encoded; } std::vector<byte> decode(const std::string& input) { if(input.length() % 4) throw std::runtime_error("Invalid base64 length!"); std::size_t padding{}; if(input.length()) { if(input[input.length() - 1] == kPadCharacter) padding++; if(input[input.length() - 2] == kPadCharacter) padding++; } std::vector<byte> decoded; decoded.reserve(((input.length() / 4) * 3) - padding); std::uint32_t temp{}; auto it = input.begin(); while(it < input.end()) { for(std::size_t i = 0; i < 4; ++i) { temp <<= 6; if (*it >= 0x41 && *it <= 0x5A) temp |= *it - 0x41; else if(*it >= 0x61 && *it <= 0x7A) temp |= *it - 0x47; else if(*it >= 0x30 && *it <= 0x39) temp |= *it + 0x04; else if(*it == 0x2B) temp |= 0x3E; else if(*it == 0x2F) temp |= 0x3F; else if(*it == kPadCharacter) { switch(input.end() - it) { case 1: decoded.push_back((temp >> 16) & 0x000000FF); decoded.push_back((temp >> 8 ) & 0x000000FF); return decoded; case 2: decoded.push_back((temp >> 10) & 0x000000FF); return decoded; default: throw std::runtime_error("Invalid padding in base64!"); } } else throw std::runtime_error("Invalid character in base64!"); ++it; } decoded.push_back((temp >> 16) & 0x000000FF); decoded.push_back((temp >> 8 ) & 0x000000FF); decoded.push_back((temp ) & 0x000000FF); } return decoded; } } |
nice!
Why not use Base91? Also, I see little point in a restructuring of the wiki code, especially as in the process you have removed the comments. Creating another variant of something already published & tested goes against the principle of software reuse. Perhaps you were bored?
I put it in a namespace. Gave static constants better names and properly marked them inline since they’re in header file. I marked the functions inline per ODR. I changed the windows specifit DWORD to a standard C++ type. I removed macros and unnececary support for unicode; ASCII is already a subset of it anyways. Comments were mostly useless after my changes so I rid of them. I see those changes as improvements to be honest. Now the code is standard conformant and portable.
Another note on the removal of wchar_t support… since the output of base64::encode contains
ONLY ASCII
characters, it can be safely converted to std::wstring like this:
or once could also use std::wstringstream and
<<
the
std::string
into the stream.
In base64.cpp you use #include “ascii_escape_code.hpp” but I cannot find it anywhere. Is it required?
My apologies! I renamed the file since those are ANSI not ASCII codes, it is here: https://github.com/mvorbrodt/blog/blob/master/src/ansi_escape_code.hpp
Hi, thanks for this! What would be the license of your version?
All the code on my blog is under 0-BSD license. Meaning you can do whatever you like with it, even include in a commercial produce without open sourceing or giving my any credit. Enjoy!