Base64 decode data

From Rosetta Code
Task
Base64 decode data
You are encouraged to solve this task according to the task description, using any language you may know.

See Base64 encode data.

Now write a program that takes the output of the Base64 encode data task as input and regenerate the original file.

When working on the VBA implementation I found several 'solutions' on the net, including one from the software maker himself, that showed output with incorrect padding. Obviously with incorrect padding in the output you can not decode correctly to the original file again.

C++[edit]

#include <algorithm>
#include <iostream>
#include <string>
#include <vector>
 
typedef unsigned char ubyte;
const auto BASE64 = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
 
std::vector<ubyte> encode(const std::vector<ubyte>& source) {
auto it = source.cbegin();
auto end = source.cend();
 
std::vector<ubyte> sink;
while (it != end) {
auto b1 = *it++;
int acc;
 
sink.push_back(BASE64[b1 >> 2]); // first output (first six bits from b1)
 
acc = (b1 & 0x3) << 4; // last two bits from b1
if (it != end) {
auto b2 = *it++;
acc |= (b2 >> 4); // first four bits from b2
 
sink.push_back(BASE64[acc]); // second output
 
acc = (b2 & 0xF) << 2; // last four bits from b2
if (it != end) {
auto b3 = *it++;
acc |= (b3 >> 6); // first two bits from b3
 
sink.push_back(BASE64[acc]); // third output
sink.push_back(BASE64[b3 & 0x3F]); // fouth output (final six bits from b3)
} else {
sink.push_back(BASE64[acc]); // third output
sink.push_back('='); // fourth output (1 byte padding)
}
} else {
sink.push_back(BASE64[acc]); // second output
sink.push_back('='); // third output (first padding byte)
sink.push_back('='); // fourth output (second padding byte)
}
}
return sink;
}
 
int findIndex(ubyte val) {
if ('A' <= val && val <= 'Z') {
return val - 'A';
}
if ('a' <= val && val <= 'z') {
return val - 'a' + 26;
}
if ('0' <= val && val <= '9') {
return val - '0' + 52;
}
if ('+' == val) {
return 62;
}
if ('/' == val) {
return 63;
}
return -1;
}
 
std::vector<ubyte> decode(const std::vector<ubyte>& source) {
if (source.size() % 4 != 0) {
throw new std::runtime_error("Error in size to the decode method");
}
 
auto it = source.cbegin();
auto end = source.cend();
 
std::vector<ubyte> sink;
while (it != end) {
auto b1 = *it++;
auto b2 = *it++;
auto b3 = *it++; // might be first padding byte
auto b4 = *it++; // might be first or second padding byte
 
auto i1 = findIndex(b1);
auto i2 = findIndex(b2);
int acc;
 
acc = i1 << 2; // six bits came from the first byte
acc |= i2 >> 4; // two bits came from the first byte
 
sink.push_back(acc); // output the first byte
 
if (b3 != '=') {
auto i3 = findIndex(b3);
 
acc = (i2 & 0xF) << 4; // four bits came from the second byte
acc |= i3 >> 2; // four bits came from the second byte
 
sink.push_back(acc); // output the second byte
 
if (b4 != '=') {
auto i4 = findIndex(b4);
 
acc = (i3 & 0x3) << 6; // two bits came from the third byte
acc |= i4; // six bits came from the third byte
 
sink.push_back(acc); // output the third byte
}
}
}
return sink;
}
 
int main() {
using namespace std;
 
string data = "VG8gZXJyIGlzIGh1bWFuLCBidXQgdG8gcmVhbGx5IGZvdWwgdGhpbmdzIHVwIHlvdSBuZWVkIGEgY29tcHV0ZXIuCiAgICAtLVBhdWwgUi5FaHJsaWNo";
vector<ubyte> datav{ begin(data), end(data) };
cout << data << "\n\n";
 
auto decoded = decode(datav);
std::for_each(cbegin(decoded), cend(decoded), [](char c) { cout << c; });
cout << '\n';
 
return 0;
}
Output:
VG8gZXJyIGlzIGh1bWFuLCBidXQgdG8gcmVhbGx5IGZvdWwgdGhpbmdzIHVwIHlvdSBuZWVkIGEgY29tcHV0ZXIuCiAgICAtLVBhdWwgUi5FaHJsaWNo

To err is human, but to really foul things up you need a computer.
    --Paul R.Ehrlich

D[edit]

Translation of: Perl 6
import std.base64;
import std.stdio;
 
void main() {
auto data = "VG8gZXJyIGlzIGh1bWFuLCBidXQgdG8gcmVhbGx5IGZvdWwgdGhpbmdzIHVwIHlvdSBuZWVkIGEgY29tcHV0ZXIuCiAgICAtLSBQYXVsIFIuIEVocmxpY2g=";
writeln(data);
writeln;
 
auto decoded = cast(char[])Base64.decode(data);
writeln(decoded);
}
Output:
VG8gZXJyIGlzIGh1bWFuLCBidXQgdG8gcmVhbGx5IGZvdWwgdGhpbmdzIHVwIHlvdSBuZWVkIGEgY29tcHV0ZXIuCiAgICAtLSBQYXVsIFIuIEVocmxpY2g=

To err is human, but to really foul things up you need a computer.
    -- Paul R. Ehrlich

Go[edit]

As images can no longer be uploaded to RC, I've encoded and decoded a string rather than the Rosetta Code icon.

package main
 
import (
"encoding/base64"
"fmt"
)
 
func main() {
msg := "Rosetta Code Base64 decode data task"
fmt.Println("Original :", msg)
encoded := base64.StdEncoding.EncodeToString([]byte(msg))
fmt.Println("\nEncoded  :", encoded)
decoded, err := base64.StdEncoding.DecodeString(encoded)
if err != nil {
fmt.Println(err)
return
}
fmt.Println("\nDecoded  :", string(decoded))
}
Output:
Original : Rosetta Code Base64 decode data task

Encoded  : Um9zZXR0YSBDb2RlIEJhc2U2NCBkZWNvZGUgZGF0YSB0YXNr

Decoded  : Rosetta Code Base64 decode data task

Jsish[edit]

See Base64_encode_data#Jsish for base64.jsi.

/* Base64 decode, in Jsish */
var data = exec('jsish base64.jsi', {retAll:true}).data; // or use File.read('stdin');
var icon = Util.base64(data, true);
File.write('rosetta-favicon.ico', icon);


Julia[edit]

Using an IOBuffer here, though not really needed to decode a string, shows how we could pipe a network stream or file though Julia's builtin Base64 decoder.

using Base64
 
io = IOBuffer()
 
iob64_decode = Base64DecodePipe(io)
 
write(io, "VG8gZXJyIGlzIGh1bWFuLCBidXQgdG8gcmVhbGx5IGZvdWwgdGhpbmdzIHVwIHlvdSBuZWVkIGEgY29tcHV0ZXIuCiAgICAtLVBhdWwgUi5FaHJsaWNo")
 
seekstart(io)
 
println(String(read(iob64_decode)))
 
Output:
To err is human, but to really foul things up you need a computer.
    --Paul R.Ehrlich

Kotlin[edit]

Translation of: D
import java.util.Base64
 
fun main() {
val data =
"VG8gZXJyIGlzIGh1bWFuLCBidXQgdG8gcmVhbGx5IGZvdWwgdGhpbmdzIHVwIHlvdSBuZWVkIGEgY29tcHV0ZXIuCiAgICAtLSBQYXVsIFIuIEVocmxpY2g="
val decoder = Base64.getDecoder()
val decoded = decoder.decode(data)
val decodedStr = String(decoded, Charsets.UTF_8)
println(decodedStr)
}
Output:
To err is human, but to really foul things up you need a computer.
    -- Paul R. Ehrlich

Perl[edit]

The MIME::Base64 module is to be preferred, but this works too.

sub decode_base64 {
my($d) = @_;
$d =~ tr!A-Za-z0-9+/!!cd;
$d =~ s/=+$//;
$d =~ tr!A-Za-z0-9+/! -_!;
my $r = '';
while( $d =~ /(.{1,60})/gs ){
my $len = chr(32 + length($1)*3/4);
$r .= unpack("u", $len . $1 );
}
$r;
}
 
$data = <<EOD;
J1R3YXMgYnJpbGxpZywgYW5kIHRoZSBzbGl0aHkgdG92ZXMKRGlkIGd5cmUgYW5kIGdpbWJsZSBp
biB0aGUgd2FiZToKQWxsIG1pbXN5IHdlcmUgdGhlIGJvcm9nb3ZlcywKQW5kIHRoZSBtb21lIHJh
dGhzIG91dGdyYWJlLgo=
EOD

 
print decode_base64($data) . "\n";
Output:
'Twas brillig, and the slithy toves
Did gyre and gimble in the wabe:
All mimsy were the borogoves,
And the mome raths outgrabe.

Perl 6[edit]

Works with: Rakudo version 2018.11
my $e64 = '
VG8gZXJyIGlzIGh1bWFuLCBidXQgdG8gcmVhbGx5IGZvdWwgdGhpbmdzIHVwIHlvdSBuZWVkIGEgY2
9tcHV0ZXIuCiAgICAtLSBQYXVsIFIuIEVocmxpY2g=
'
;
 
my @base64map = flat 'A' .. 'Z', 'a' .. 'z', ^10, '+', '/';
my %base64 is default(0) = @base64map.pairs.invert;
 
sub base64-decode-slow ($enc) {
my $buf = Buf.new;
for $enc.subst(/\s/, '', :g).comb(4) -> $chunck {
$buf.append: |(sprintf "%06d%06d%06d%06d", |$chunck.comb.map:
{%base64{$_}.base(2)}).comb(8).map: {:2($_)};
}
$buf
}
 
say 'Slow:';
say base64-decode-slow($e64).decode('utf8');
 
 
# Of course, the above routine is slow and is only for demonstration purposes.
# For real code you should use a module, which is MUCH faster and heavily tested.
say "\nFast:";
use Base64::Native;
say base64-decode($e64).decode('utf8');
Output:
Slow:
To err is human, but to really foul things up you need a computer.
    -- Paul R. Ehrlich

Fast:
To err is human, but to really foul things up you need a computer.
    -- Paul R. Ehrlich

Phix[edit]

include builtins\base64.e
string s = "Rosetta Code Base64 decode data task"
string e = encode_base64(s)
?e
?decode_base64(e)
Output:
"Um9zZXR0YSBDb2RlIEJhc2U2NCBkZWNvZGUgZGF0YSB0YXNr"
"Rosetta Code Base64 decode data task"

Ruby[edit]

require 'base64'
 
perl6_example ='
VG8gZXJyIGlzIGh1bWFuLCBidXQgdG8gcmVhbGx5IGZvdWwgdGhpbmdzIHVwIHlvdSBuZWVkIGEgY2
9tcHV0ZXIuCiAgICAtLSBQYXVsIFIuIEVocmxpY2g=
'

puts Base64.decode64 perl6_example
Output:
To err is human, but to really foul things up you need a computer.
    -- Paul R. Ehrlich

Seed7[edit]

The Seed7 library encoding.s7i defines the functions toBase64 and fromBase64.

$ include "seed7_05.s7i";
include "gethttp.s7i";
include "encoding.s7i";
 
const proc: main is func
local
var string: original is "";
var string: encoded is "";
begin
original := getHttp("rosettacode.org/favicon.ico");
encoded := toBase64(original);
writeln("Is the Rosetta Code icon the same (byte for byte) encoded then decoded: " <&
fromBase64(encoded) = original);
end func;
Output:
Is the Rosetta Code icon the same (byte for byte) encoded then decoded: TRUE

Sidef[edit]

var data = <<'EOT'
VG8gZXJyIGlzIGh1bWFuLCBidXQgdG8gcmVhbGx5IGZvdWwgdGhpbmdzIHVwIHlvdSBuZWVkIGEgY2
9tcHV0ZXIuCiAgICAtLSBQYXVsIFIuIEVocmxpY2g=
EOT
 
say data.decode_base64
Output:
To err is human, but to really foul things up you need a computer.
    -- Paul R. Ehrlich

zkl[edit]

Using shared libraries for cURL and message hashing:

var [const] MsgHash=Import("zklMsgHash"), Curl=Import("zklCurl");
 
icon:=Curl().get("http://rosettacode.org/favicon.ico"); //-->(Data(4,331),693,0)
icon=icon[0][icon[1],*]; // remove header
iconEncoded:=MsgHash.base64encode(icon);
iconDecoded:=MsgHash.base64decode(iconEncoded);
File("rosettaCodeIcon.ico","wb").write(iconDecoded); # eyeball checking says good
println("Is the Rosetta Code icon the same (byte for byte) encoded then decoded: ",
icon==iconDecoded);
Output:
Is the Rosetta Code icon the same (byte for byte) encoded then decoded: True
Text based test:
msg,b64 := "Rosetta Code Base64 decode data task", MsgHash.base64encode(msg);
println("Original: %s\nEncoded:  %s\nBytes:  %s\nDecoded:  %s"
.fmt(msg, b64.text, b64.bytes().apply("toString",16).concat(","),
MsgHash.base64decode(b64).text));
Original: Rosetta Code Base64 decode data task
Encoded:  Um9zZXR0YSBDb2RlIEJhc2U2NCBkZWNvZGUgZGF0YSB0YXNr

Bytes:    55,6d,39,7a,5a,58,52,30,59,53,42,44,62,32,52,6c,49,45,4a,68,63,32,55,32,4e,43,42,6b,5a,57,4e,76,5a,47,55,67,5a,47,46,30,59,53,42,30,59,58,4e,72,a
Decoded:  Rosetta Code Base64 decode data task