RandomTest.php

Go to the documentation of this file.

<?php
if( PHP_SAPI != 'cli' ) {
    die( "Run me from the command line please.\n" );
}
 
require_once 'UtfNormal.php';
require_once '../diff/DifferenceEngine.php';
 
dl( 'php_utfnormal.so' );
 
# mt_srand( 99999 );
 
function randomString( $length, $nullOk, $ascii = false ) {
    $out = '';
    for( $i = 0; $i < $length; $i++ )
        $out .= chr( mt_rand( $nullOk ? 0 : 1, $ascii ? 127 : 255 ) );
    return $out;
}
 
/* Duplicate of the cleanUp() path for ICU usage */
function donorm( $str ) {
    # We exclude a few chars that ICU would not.
    $str = preg_replace( '/[\x00-\x08\x0b\x0c\x0e-\x1f]/', UTF8_REPLACEMENT, $str );
    $str = str_replace( UTF8_FFFE, UTF8_REPLACEMENT, $str );
    $str = str_replace( UTF8_FFFF, UTF8_REPLACEMENT, $str );
 
    # UnicodeString constructor fails if the string ends with a head byte.
    # Add a junk char at the end, we'll strip it off
    return rtrim( utf8_normalize( $str . "\x01", UtfNormal::UNORM_NFC ), "\x01" );
}
 
function showDiffs( $a, $b ) {
    $ota = explode( "\n", str_replace( "\r\n", "\n", $a ) );
    $nta = explode( "\n", str_replace( "\r\n", "\n", $b ) );
 
    $diffs = new Diff( $ota, $nta );
    $formatter = new TableDiffFormatter();
    $funky = $formatter->format( $diffs );
    $matches = array();
    preg_match_all( '/<(?:ins|del) class="diffchange">(.*?)<\/(?:ins|del)>/', $funky, $matches );
    foreach( $matches[1] as $bit ) {
        $hex = bin2hex( $bit );
        echo "\t$hex\n";
    }
}
 
$size = 16;
$n = 0;
while( true ) {
    $n++;
    echo "$n\n";
 
    $str = randomString( $size, true);
    $clean = UtfNormal::cleanUp( $str );
    $norm = donorm( $str );
 
    echo strlen( $clean ) . ", " . strlen( $norm );
    if( $clean == $norm ) {
        echo " (match)\n";
    } else {
        echo " (FAIL)\n";
        echo "\traw: " . bin2hex( $str ) . "\n" .
             "\tphp: " . bin2hex( $clean ) . "\n" .
             "\ticu: " . bin2hex( $norm ) . "\n";
        echo "\n\tdiffs:\n";
        showDiffs( $clean, $norm );
        die();
    }
 
 
    $str = '';
    $clean = '';
    $norm = '';
}