includes/normal/UtfNormalTest.php

   1 <?php
   2 /**
   3  * Implements the conformance test at:
   4  * http://www.unicode.org/Public/UNIDATA/NormalizationTest.txt
   5  *
   6  * Copyright © 2004 Brion Vibber <brion@pobox.com>
   7  * http://www.mediawiki.org/
   8  *
   9  * This program is free software; you can redistribute it and/or modify
  10  * it under the terms of the GNU General Public License as published by
  11  * the Free Software Foundation; either version 2 of the License, or
  12  * (at your option) any later version.
  13  *
  14  * This program is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  17  * GNU General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU General Public License along
  20  * with this program; if not, write to the Free Software Foundation, Inc.,
  21  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  22  * http://www.gnu.org/copyleft/gpl.html
  23  *
  24  * @file
  25  * @ingroup UtfNormal
  26  */
  27
  28 if( PHP_SAPI != 'cli' ) {
  29         die( "Run me from the command line please.\n" );
  30 }
  31
  32 $verbose = true;
  33 #define( 'PRETTY_UTF8', true );
  34
  35 if( defined( 'PRETTY_UTF8' ) ) {
  36         function pretty( $string ) {
  37                 return strtoupper( bin2hex( $string ) );
  38         }
  39 } else {
  40         /**
  41          * @ignore
  42          * @return string
  43          */
  44         function pretty( $string ) {
  45                 return strtoupper( utf8ToHexSequence( $string ) );
  46         }
  47 }
  48
  49 if( isset( $_SERVER['argv'] ) && in_array( '--icu', $_SERVER['argv'] ) ) {
  50         dl( 'php_utfnormal.so' );
  51 }
  52
  53 require_once 'UtfNormalDefines.php';
  54 require_once 'UtfNormalUtil.php';
  55 require_once 'UtfNormal.php';
  56
  57 $in = fopen("NormalizationTest.txt", "rt");
  58 if( !$in ) {
  59         print "Couldn't open NormalizationTest.txt -- can't run tests.\n";
  60         print "If necessary, manually download this file. It can be obtained at\n";
  61         print "http://www.unicode.org/Public/UNIDATA/NormalizationTest.txt";
  62         exit(-1);
  63 }
  64
  65 $normalizer = new UtfNormal;
  66
  67 $total = 0;
  68 $success = 0;
  69 $failure = 0;
  70 $ok = true;
  71 $testedChars = array();
  72 while( false !== ( $line = fgets( $in ) ) ) {
  73         list( $data, $comment ) = explode( '#', $line );
  74         if( $data === '' ) continue;
  75         $matches = array();
  76         if( preg_match( '/@Part([\d])/', $data, $matches ) ) {
  77                 if( $matches[1] > 0 ) {
  78                         $ok = reportResults( $total, $success, $failure ) && $ok;
  79                 }
  80                 print "Part {$matches[1]}: $comment";
  81                 continue;
  82         }
  83
  84         $columns = array_map( "hexSequenceToUtf8", explode( ";", $data ) );
  85         array_unshift( $columns, '' );
  86
  87         $testedChars[$columns[1]] = true;
  88         $total++;
  89         if( testNormals( $normalizer, $columns, $comment, $verbose ) ) {
  90                 $success++;
  91         } else {
  92                 $failure++;
  93                 # print "FAILED: $comment";
  94         }
  95         if( $total % 100 == 0 ) print "$total ";
  96 }
  97 fclose( $in );
  98
  99 $ok = reportResults( $total, $success, $failure ) && $ok;
 100
 101 $in = fopen("UnicodeData.txt", "rt" );
 102 if( !$in ) {
 103         print "Can't open UnicodeData.txt for reading.\n";
 104         print "If necessary, fetch this file from the internet:\n";
 105         print "http://www.unicode.org/Public/UNIDATA/UnicodeData.txt\n";
 106         exit(-1);
 107 }
 108 print "Now testing invariants...\n";
 109 while( false !== ($line = fgets( $in ) ) ) {
 110         $cols = explode( ';', $line );
 111         $char = codepointToUtf8( hexdec( $cols[0] ) );
 112         $desc = $cols[0] . ": " . $cols[1];
 113         if( $char < "\x20" || $char >= UTF8_SURROGATE_FIRST && $char <= UTF8_SURROGATE_LAST ) {
 114                 # Can't check NULL with the ICU plugin, as null bytes fail in C land.
 115                 # Skip other control characters, as we strip them for XML safety.
 116                 # Surrogates are illegal on their own or in UTF-8, ignore.
 117                 continue;
 118         }
 119         if( empty( $testedChars[$char] ) ) {
 120                 $total++;
 121                 if( testInvariant( $normalizer, $char, $desc, $verbose ) ) {
 122                         $success++;
 123                 } else {
 124                         $failure++;
 125                 }
 126                 if( $total % 100 == 0 ) print "$total ";
 127         }
 128 }
 129 fclose( $in );
 130
 131 $ok = reportResults( $total, $success, $failure ) && $ok;
 132
 133 if( $ok ) {
 134         print "TEST SUCCEEDED!\n";
 135         exit(0);
 136 } else {
 137         print "TEST FAILED!\n";
 138         exit(-1);
 139 }
 140
 141 ## ------
 142
 143 function reportResults( &$total, &$success, &$failure ) {
 144         $percSucc = intval( $success * 100 / $total );
 145         $percFail = intval( $failure * 100 / $total );
 146         print "\n";
 147         print "$success tests successful ($percSucc%)\n";
 148         print "$failure tests failed ($percFail%)\n\n";
 149         $ok = ($success > 0 && $failure == 0);
 150         $total = 0;
 151         $success = 0;
 152         $failure = 0;
 153         return $ok;
 154 }
 155
 156 function testNormals( &$u, $c, $comment, $verbose, $reportFailure = false ) {
 157         $result = testNFC( $u, $c, $comment, $reportFailure );
 158         $result = testNFD( $u, $c, $comment, $reportFailure ) && $result;
 159         $result = testNFKC( $u, $c, $comment, $reportFailure ) && $result;
 160         $result = testNFKD( $u, $c, $comment, $reportFailure ) && $result;
 161         $result = testCleanUp( $u, $c, $comment, $reportFailure ) && $result;
 162
 163         if( $verbose && !$result && !$reportFailure ) {
 164                 print $comment;
 165                 testNormals( $u, $c, $comment, $verbose, true );
 166         }
 167         return $result;
 168 }
 169
 170 function verbosify( $a, $b, $col, $form, $verbose ) {
 171         #$result = ($a === $b);
 172         $result = (strcmp( $a, $b ) == 0);
 173         if( $verbose ) {
 174                 $aa = pretty( $a );
 175                 $bb = pretty( $b );
 176                 $ok = $result ? "succeed" : " failed";
 177                 $eq = $result ? "==" : "!=";
 178                 print "  $ok $form c$col '$aa' $eq '$bb'\n";
 179         }
 180         return $result;
 181 }
 182
 183 function testNFC( &$u, $c, $comment, $verbose ) {
 184         $result = verbosify( $c[2], $u->toNFC( $c[1] ), 1, 'NFC', $verbose );
 185         $result = verbosify( $c[2], $u->toNFC( $c[2] ), 2, 'NFC', $verbose ) && $result;
 186         $result = verbosify( $c[2], $u->toNFC( $c[3] ), 3, 'NFC', $verbose ) && $result;
 187         $result = verbosify( $c[4], $u->toNFC( $c[4] ), 4, 'NFC', $verbose ) && $result;
 188         $result = verbosify( $c[4], $u->toNFC( $c[5] ), 5, 'NFC', $verbose ) && $result;
 189         return $result;
 190 }
 191
 192 function testCleanUp( &$u, $c, $comment, $verbose ) {
 193         $x = $c[1];
 194         $result = verbosify( $c[2], $u->cleanUp( $x ), 1, 'cleanUp', $verbose );
 195         $x = $c[2];
 196         $result = verbosify( $c[2], $u->cleanUp( $x ), 2, 'cleanUp', $verbose ) && $result;
 197         $x = $c[3];
 198         $result = verbosify( $c[2], $u->cleanUp( $x ), 3, 'cleanUp', $verbose ) && $result;
 199         $x = $c[4];
 200         $result = verbosify( $c[4], $u->cleanUp( $x ), 4, 'cleanUp', $verbose ) && $result;
 201         $x = $c[5];
 202         $result = verbosify( $c[4], $u->cleanUp( $x ), 5, 'cleanUp', $verbose ) && $result;
 203         return $result;
 204 }
 205
 206 function testNFD( &$u, $c, $comment, $verbose ) {
 207         $result = verbosify( $c[3], $u->toNFD( $c[1] ), 1, 'NFD', $verbose );
 208         $result = verbosify( $c[3], $u->toNFD( $c[2] ), 2, 'NFD', $verbose ) && $result;
 209         $result = verbosify( $c[3], $u->toNFD( $c[3] ), 3, 'NFD', $verbose ) && $result;
 210         $result = verbosify( $c[5], $u->toNFD( $c[4] ), 4, 'NFD', $verbose ) && $result;
 211         $result = verbosify( $c[5], $u->toNFD( $c[5] ), 5, 'NFD', $verbose ) && $result;
 212         return $result;
 213 }
 214
 215 function testNFKC( &$u, $c, $comment, $verbose ) {
 216         $result = verbosify( $c[4], $u->toNFKC( $c[1] ), 1, 'NFKC', $verbose );
 217         $result = verbosify( $c[4], $u->toNFKC( $c[2] ), 2, 'NFKC', $verbose ) && $result;
 218         $result = verbosify( $c[4], $u->toNFKC( $c[3] ), 3, 'NFKC', $verbose ) && $result;
 219         $result = verbosify( $c[4], $u->toNFKC( $c[4] ), 4, 'NFKC', $verbose ) && $result;
 220         $result = verbosify( $c[4], $u->toNFKC( $c[5] ), 5, 'NFKC', $verbose ) && $result;
 221         return $result;
 222 }
 223
 224 function testNFKD( &$u, $c, $comment, $verbose ) {
 225         $result = verbosify( $c[5], $u->toNFKD( $c[1] ), 1, 'NFKD', $verbose );
 226         $result = verbosify( $c[5], $u->toNFKD( $c[2] ), 2, 'NFKD', $verbose ) && $result;
 227         $result = verbosify( $c[5], $u->toNFKD( $c[3] ), 3, 'NFKD', $verbose ) && $result;
 228         $result = verbosify( $c[5], $u->toNFKD( $c[4] ), 4, 'NFKD', $verbose ) && $result;
 229         $result = verbosify( $c[5], $u->toNFKD( $c[5] ), 5, 'NFKD', $verbose ) && $result;
 230         return $result;
 231 }
 232
 233 function testInvariant( &$u, $char, $desc, $verbose, $reportFailure = false ) {
 234         $result = verbosify( $char, $u->toNFC( $char ), 1, 'NFC', $reportFailure );
 235         $result = verbosify( $char, $u->toNFD( $char ), 1, 'NFD', $reportFailure ) && $result;
 236         $result = verbosify( $char, $u->toNFKC( $char ), 1, 'NFKC', $reportFailure ) && $result;
 237         $result = verbosify( $char, $u->toNFKD( $char ), 1, 'NFKD', $reportFailure ) && $result;
 238         $result = verbosify( $char, $u->cleanUp( $char ), 1, 'cleanUp', $reportFailure ) && $result;
 239
 240         if( $verbose && !$result && !$reportFailure ) {
 241                 print $desc;
 242                 testInvariant( $u, $char, $desc, $verbose, true );
 243         }
 244         return $result;
 245 }