Testing Gettext MO hash table compilation in PHP

abc.mo

abc.po

# Compile this to abc.mo to run mo-hash.php
# $ msgfmt abc.po -o abc.mo
# 
msgid ""
msgstr ""
"Project-Id-Version: Test\n"
"Language: English\n"
"Plural-Forms: nplurals=2; plural=n!=1\n"
"MIME-Version: 1.0\n"
"Content-Type: text/plain; charset=UTF-8\n"
"Content-Transfer-Encoding: 8bit"

msgid "a"
msgstr "X"

msgid "b"
msgstr "Y"

msgid "c"
msgstr "Z"

mo-hash.php

<?php
/**
 * MO hash table test.
 */
 
 
// Compile hash table from simple array including empty header key. 
$msgids = array( '', 'a', 'b', 'c' );
$table = compile_table( $msgids );

echo "Hash table compiled by PHP: ";
print_binary( $table );


// Pull hash table from abc.mo compiled by gettext msgfmt
$mo = file_get_contents('abc.mo');
list(,$magic) = unpack('V', substr($mo,0,4) );
$fmt = 0x950412de === $magic ? 'V*' : ( 0xde120495 === $magic ? 'N' : die('Bad MO file') );
list(,$s,$h) = unpack( $fmt, substr( $mo, 20, 8 ) );
$table = substr( $mo, $h, $s * 4 );

echo "Hash table from abc.mo: ";
print_binary( $table );



/**
 * Displays hash table as hex
 */
function print_binary( $str ){
   $pairs = str_split( bin2hex($str), 2 );
   $lines = str_split( implode(' ', $pairs ), 12 );
   echo strlen($str)," bytes\n",implode("\n", $lines), "\n\n";
}
 
 

/**
 * Build hash table string from array of msgid strings. (From write-mo.c)
 * http://fossies.org/dox/gettext-0.18.3.1/write-mo_8c_source.html#l00384
 */ 
function compile_table( array $msgids ){
    $j = 0;
    $n = count($msgids);
    $nul_word = "\0\0\0\0";
    $hash_tab_size = max( 3, next_prime( $n * 4 / 3 ) );
    $map = array_fill( 0, $hash_tab_size, $nul_word );
    foreach( $msgids as $msgid ){
        $hash_val = hash_string( $msgid );
        $idx = $hash_val % $hash_tab_size;
        // We need the second hashing function if another item in this slot already
        if( $nul_word !== $map[$idx] ){
            $incr = 1 + ($hash_val % ( $hash_tab_size - 2 ) );
            do {
                $diff = $hash_tab_size - $incr;
                if ( $idx >= $diff ) {
                    $idx -= $diff;
                }
                else {
                    $idx += $incr;
                }
            }
            while( $nul_word !== $map[$idx] );
        }
        $map[$idx] = pack( 'V', ++$j );
    }
    return implode('',$map);
} 
  
  

/**
 * hashpjw function by P.J. Weinberger. (From hash-string.c)
 * http://fossies.org/dox/gettext-0.18.3.1/hash-string_8c_source.html#l00031
 */
function hash_string( $str ){
    $i = -1;
    $hval = 0;
    $len = strlen($str);
    while( ++$i < $len ){
        $ord = ord( $str{$i} );
        $hval = ( $hval << 4 ) + $ord;
        $g = $hval & 0xf0000000;
        if( $g !== 0 ){
            $hval ^= $g >> 24;
            $hval ^= $g;
        }
    }
    return $hval;
}



/**
 * Get next prime number up from seed number. (From hash.c)
 * http://fossies.org/dox/gettext-0.18.3.1/gettext-tools_2gnulib-lib_2hash_8c_source.html#l00073
 */
function next_prime( $seed ){
    $seed |= 1;
    while ( ! is_prime($seed) ){
        $seed += 2;
    }
    return $seed;
}



/**
 * Test if a given integer is a prime number.
 * http://icdif.com/computing/2011/09/15/check-number-prime-number/
 */
function is_prime( $num ) {
    if ($num === 1 ){
        return false;
    }
    if( $num === 2 ){
        return true;
    }
    if( $num % 2 === 0 ) {
        return false;
    }
    $max = ceil( sqrt($num) );
    for( $i = 3; $i <= $max; $i+=2 ) {
        if( $num % $i === 0 ){
            return false;
        }
    }
    return true;
}