effective solution: base32 encoding in php

Question

I am looking for a base32 function/class for php. the different classes and function that i found are all very ineffective. I ran a benchmark and came to the following resul

Answer 1

You can try these functions I adapted from bbars and crockford:

function crockford32_encode($data) {
    $chars = '0123456789abcdefghjkmnpqrstvwxyz';
    $mask = 0b11111;

    $dataSize = strlen($data);
    $res = '';
    $remainder = 0;
    $remainderSize = 0;

    for($i = 0; $i < $dataSize; $i++) {
        $b = ord($data[$i]);
        $remainder = ($remainder << 8) | $b;
        $remainderSize += 8;
        while($remainderSize > 4) {
            $remainderSize -= 5;
            $c = $remainder & ($mask << $remainderSize);
            $c >>= $remainderSize;
            $res .= $chars[$c];
        }
    }
    if($remainderSize > 0) {
        $remainder <<= (5 - $remainderSize);
        $c = $remainder & $mask;
        $res .= $chars[$c];
    }

    return $res;
}

function crockford32_decode($data) {
    $map = [
        '0' => 0,
        'O' => 0,
        'o' => 0,
        '1' => 1,
        'I' => 1,
        'i' => 1,
        'L' => 1,
        'l' => 1,
        '2' => 2,
        '3' => 3,
        '4' => 4,
        '5' => 5,
        '6' => 6,
        '7' => 7,
        '8' => 8,
        '9' => 9,
        'A' => 10,
        'a' => 10,
        'B' => 11,
        'b' => 11,
        'C' => 12,
        'c' => 12,
        'D' => 13,
        'd' => 13,
        'E' => 14,
        'e' => 14,
        'F' => 15,
        'f' => 15,
        'G' => 16,
        'g' => 16,
        'H' => 17,
        'h' => 17,
        'J' => 18,
        'j' => 18,
        'K' => 19,
        'k' => 19,
        'M' => 20,
        'm' => 20,
        'N' => 21,
        'n' => 21,
        'P' => 22,
        'p' => 22,
        'Q' => 23,
        'q' => 23,
        'R' => 24,
        'r' => 24,
        'S' => 25,
        's' => 25,
        'T' => 26,
        't' => 26,
        'V' => 27,
        'v' => 27,
        'W' => 28,
        'w' => 28,
        'X' => 29,
        'x' => 29,
        'Y' => 30,
        'y' => 30,
        'Z' => 31,
        'z' => 31,
    ];

    $data = strtolower($data);
    $dataSize = strlen($data);
    $buf = 0;
    $bufSize = 0;
    $res = '';

    for($i = 0; $i < $dataSize; $i++) {
        $c = $data[$i];
        if(!isset($map[$c])) {
            throw new \Exception("Unsupported character $c (0x".bin2hex($c).") at position $i");
        }
        $b = $map[$c];
        $buf = ($buf << 5) | $b;
        $bufSize += 5;
        if($bufSize > 7) {
            $bufSize -= 8;
            $b = ($buf & (0xff << $bufSize)) >> $bufSize;
            $res .= chr($b);
        }
    }

    return $res;
}

Answer 2

Try this: https://github.com/bbars/utils/blob/master/php-base32-encode-decode/Base32.php

Uses case-insesitive alpabet [0-9, a-v] and works faster than Base2n(5):

size   | Base32::encode  | Base32::decode  | $base2n->encode  | $base2n->decode
-------------------------------------------------------------------------------
1      | 0.0000331401825 | 0.0000088214874 | 0.0002369880676  | 0.0001671314240
2      | 0.0000050067902 | 0.0000040531158 | 0.0000100135803  | 0.0000081062317
4      | 0.0000050067902 | 0.0000059604645 | 0.0000097751617  | 0.0000100135803
8      | 0.0000078678131 | 0.0000100135803 | 0.0000131130219  | 0.0000140666962
16     | 0.0000128746033 | 0.0000178813934 | 0.0000250339508  | 0.0000250339508
32     | 0.0000238418579 | 0.0000319480896 | 0.0000441074371  | 0.0000472068787
64     | 0.0001170635223 | 0.0000629425049 | 0.0000870227814  | 0.0000259876251
128    | 0.0000879764557 | 0.0001208782196 | 0.0001959800720  | 0.0001759529114
256    | 0.0001969337463 | 0.0002408027649 | 0.0004429817200  | 0.0003459453583
512    | 0.0003631114960 | 0.0004880428314 | 0.0021460056305  | 0.0006039142609
1024   | 0.0014970302582 | 0.0009729862213 | 0.0108621120453  | 0.0015850067139
2048   | 0.0013530254364 | 0.0018491744995 | 0.0312080383301  | 0.0027630329132
4096   | 0.0027470588684 | 0.0038080215454 | 0.1312029361725  | 0.0064430236816
8192   | 0.0064270496368 | 0.0086290836334 | 0.5233020782471  | 0.0121779441833
16384  | 0.0112588405609 | 0.0167109966278 | 2.0316259860992  | 0.0277659893036
32768  | 0.0235319137573 | 0.0335960388184 | 11.6220989227295 | 0.0498571395874
65536  | 0.0478749275208 | 0.0648550987244 |                  |                
131072 | 0.1030550003052 | 0.1504058837891 |                  |                
262144 | 0.1995100975037 | 0.2654621601105 |                  |                
524288 | 0.3903131484985 | 0.5326008796692 |                  |                

Answer 3

For Base32 in PHP, you can try my implementation here:

https://github.com/ademarre/binary-to-text-php

Copied from the Base32 example in the README file:

// RFC 4648 base32 alphabet; case-insensitive
$base32 = new Base2n(5, 'ABCDEFGHIJKLMNOPQRSTUVWXYZ234567', FALSE, TRUE, TRUE);
$encoded = $base32->encode('encode this');
// MVXGG33EMUQHI2DJOM======

It's not slow, and it may or may not be faster than the class you benchmarked, but it will not be as fast as a built-in PHP function like base64_encode(). If that is very important to you, and you don't really care about Base32 encoding, then you should just use hexadecimal. You can encode hexadecimal with native PHP functions, and it is case-insensitive.

$encoded = bin2hex('encode this'); // 656e636f64652074686973
$decoded = pack('H*', $encoded);   // encode this

// Alternatively, as of PHP 5.4...
$decoded = hex2bin($encoded);      // encode this

The downside to hexadecimal is that there is more data inflation compared to Base32. Hexadecimal inflates the data 100%, while Base32 inflates the data about 60%.