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  * https://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          * @param string $string
  43          * @return string
  44          */
  45         function pretty( $string ) {
  46                 return strtoupper( utf8ToHexSequence( $string ) );
  47         }
  48 }
  49
  50 if ( isset( $_SERVER['argv'] ) && in_array( '--icu', $_SERVER['argv'] ) ) {
  51         dl( 'php_utfnormal.so' );
  52 }
  53
  54 require_once 'UtfNormalDefines.php';
  55 require_once 'UtfNormalUtil.php';
  56 require_once 'UtfNormal.php';
  57
  58 $in = fopen( "NormalizationTest.txt", "rt" );
  59 if ( !$in ) {
  60         print "Couldn't open NormalizationTest.txt -- can't run tests.\n";
  61         print "If necessary, manually download this file. It can be obtained at\n";
  62         print "http://www.unicode.org/Public/UNIDATA/NormalizationTest.txt";
  63         exit( -1 );
  64 }
  65
  66 $normalizer = new UtfNormal;
  67
  68 $total = 0;
  69 $success = 0;
  70 $failure = 0;
  71 $ok = true;
  72 $testedChars = array();
  73
  74 while ( false !== ( $line = fgets( $in ) ) ) {
  75         list( $data, $comment ) = explode( '#', $line );
  76         if ( $data === '' ) continue;
  77         $matches = array();
  78         if ( preg_match( '/@Part([\d])/', $data, $matches ) ) {
  79                 if ( $matches[1] > 0 ) {
  80                         $ok = reportResults( $total, $success, $failure ) && $ok;
  81                 }
  82                 print "Part {$matches[1]}: $comment";
  83                 continue;
  84         }
  85
  86         $columns = array_map( "hexSequenceToUtf8", explode( ";", $data ) );
  87         array_unshift( $columns, '' );
  88
  89         $testedChars[$columns[1]] = true;
  90         $total++;
  91         if ( testNormals( $normalizer, $columns, $comment, $verbose ) ) {
  92                 $success++;
  93         } else {
  94                 $failure++;
  95                 # print "FAILED: $comment";
  96         }
  97         if ( $total % 100 == 0 ) print "$total ";
  98 }
  99 fclose( $in );
 100
 101 $ok = reportResults( $total, $success, $failure ) && $ok;
 102
 103 $in = fopen( "UnicodeData.txt", "rt" );
 104 if ( !$in ) {
 105         print "Can't open UnicodeData.txt for reading.\n";
 106         print "If necessary, fetch this file from the internet:\n";
 107         print "http://www.unicode.org/Public/UNIDATA/UnicodeData.txt\n";
 108         exit( -1 );
 109 }
 110 print "Now testing invariants...\n";
 111
 112 while ( false !== ( $line = fgets( $in ) ) ) {
 113         $cols = explode( ';', $line );
 114         $char = codepointToUtf8( hexdec( $cols[0] ) );
 115         $desc = $cols[0] . ": " . $cols[1];
 116         if ( $char < "\x20" || $char >= UTF8_SURROGATE_FIRST && $char <= UTF8_SURROGATE_LAST ) {
 117                 # Can't check NULL with the ICU plugin, as null bytes fail in C land.
 118                 # Skip other control characters, as we strip them for XML safety.
 119                 # Surrogates are illegal on their own or in UTF-8, ignore.
 120                 continue;
 121         }
 122         if ( empty( $testedChars[$char] ) ) {
 123                 $total++;
 124                 if ( testInvariant( $normalizer, $char, $desc, $verbose ) ) {
 125                         $success++;
 126                 } else {
 127                         $failure++;
 128                 }
 129                 if ( $total % 100 == 0 ) print "$total ";
 130         }
 131 }
 132 fclose( $in );
 133
 134 $ok = reportResults( $total, $success, $failure ) && $ok;
 135
 136 if ( $ok ) {
 137         print "TEST SUCCEEDED!\n";
 138         exit( 0 );
 139 } else {
 140         print "TEST FAILED!\n";
 141         exit( -1 );
 142 }
 143
 144 ## ------
 145
 146 function reportResults( &$total, &$success, &$failure ) {
 147         $percSucc = intval( $success * 100 / $total );
 148         $percFail = intval( $failure * 100 / $total );
 149         print "\n";
 150         print "$success tests successful ($percSucc%)\n";
 151         print "$failure tests failed ($percFail%)\n\n";
 152         $ok = ( $success > 0 && $failure == 0 );
 153         $total = 0;
 154         $success = 0;
 155         $failure = 0;
 156
 157         return $ok;
 158 }
 159
 160 function testNormals( &$u, $c, $comment, $verbose, $reportFailure = false ) {
 161         $result = testNFC( $u, $c, $comment, $reportFailure );
 162         $result = testNFD( $u, $c, $comment, $reportFailure ) && $result;
 163         $result = testNFKC( $u, $c, $comment, $reportFailure ) && $result;
 164         $result = testNFKD( $u, $c, $comment, $reportFailure ) && $result;
 165         $result = testCleanUp( $u, $c, $comment, $reportFailure ) && $result;
 166
 167         if ( $verbose && !$result && !$reportFailure ) {
 168                 print $comment;
 169                 testNormals( $u, $c, $comment, $verbose, true );
 170         }
 171
 172         return $result;
 173 }
 174
 175 function verbosify( $a, $b, $col, $form, $verbose ) {
 176         #$result = ($a === $b);
 177         $result = ( strcmp( $a, $b ) == 0 );
 178         if ( $verbose ) {
 179                 $aa = pretty( $a );
 180                 $bb = pretty( $b );
 181                 $ok = $result ? "succeed" : " failed";
 182                 $eq = $result ? "==" : "!=";
 183                 print "  $ok $form c$col '$aa' $eq '$bb'\n";
 184         }
 185
 186         return $result;
 187 }
 188
 189 function testNFC( &$u, $c, $comment, $verbose ) {
 190         $result = verbosify( $c[2], $u->toNFC( $c[1] ), 1, 'NFC', $verbose );
 191         $result = verbosify( $c[2], $u->toNFC( $c[2] ), 2, 'NFC', $verbose ) && $result;
 192         $result = verbosify( $c[2], $u->toNFC( $c[3] ), 3, 'NFC', $verbose ) && $result;
 193         $result = verbosify( $c[4], $u->toNFC( $c[4] ), 4, 'NFC', $verbose ) && $result;
 194         $result = verbosify( $c[4], $u->toNFC( $c[5] ), 5, 'NFC', $verbose ) && $result;
 195
 196         return $result;
 197 }
 198
 199 function testCleanUp( &$u, $c, $comment, $verbose ) {
 200         $x = $c[1];
 201         $result = verbosify( $c[2], $u->cleanUp( $x ), 1, 'cleanUp', $verbose );
 202         $x = $c[2];
 203         $result = verbosify( $c[2], $u->cleanUp( $x ), 2, 'cleanUp', $verbose ) && $result;
 204         $x = $c[3];
 205         $result = verbosify( $c[2], $u->cleanUp( $x ), 3, 'cleanUp', $verbose ) && $result;
 206         $x = $c[4];
 207         $result = verbosify( $c[4], $u->cleanUp( $x ), 4, 'cleanUp', $verbose ) && $result;
 208         $x = $c[5];
 209         $result = verbosify( $c[4], $u->cleanUp( $x ), 5, 'cleanUp', $verbose ) && $result;
 210
 211         return $result;
 212 }
 213
 214 function testNFD( &$u, $c, $comment, $verbose ) {
 215         $result = verbosify( $c[3], $u->toNFD( $c[1] ), 1, 'NFD', $verbose );
 216         $result = verbosify( $c[3], $u->toNFD( $c[2] ), 2, 'NFD', $verbose ) && $result;
 217         $result = verbosify( $c[3], $u->toNFD( $c[3] ), 3, 'NFD', $verbose ) && $result;
 218         $result = verbosify( $c[5], $u->toNFD( $c[4] ), 4, 'NFD', $verbose ) && $result;
 219         $result = verbosify( $c[5], $u->toNFD( $c[5] ), 5, 'NFD', $verbose ) && $result;
 220
 221         return $result;
 222 }
 223
 224 function testNFKC( &$u, $c, $comment, $verbose ) {
 225         $result = verbosify( $c[4], $u->toNFKC( $c[1] ), 1, 'NFKC', $verbose );
 226         $result = verbosify( $c[4], $u->toNFKC( $c[2] ), 2, 'NFKC', $verbose ) && $result;
 227         $result = verbosify( $c[4], $u->toNFKC( $c[3] ), 3, 'NFKC', $verbose ) && $result;
 228         $result = verbosify( $c[4], $u->toNFKC( $c[4] ), 4, 'NFKC', $verbose ) && $result;
 229         $result = verbosify( $c[4], $u->toNFKC( $c[5] ), 5, 'NFKC', $verbose ) && $result;
 230
 231         return $result;
 232 }
 233
 234 function testNFKD( &$u, $c, $comment, $verbose ) {
 235         $result = verbosify( $c[5], $u->toNFKD( $c[1] ), 1, 'NFKD', $verbose );
 236         $result = verbosify( $c[5], $u->toNFKD( $c[2] ), 2, 'NFKD', $verbose ) && $result;
 237         $result = verbosify( $c[5], $u->toNFKD( $c[3] ), 3, 'NFKD', $verbose ) && $result;
 238         $result = verbosify( $c[5], $u->toNFKD( $c[4] ), 4, 'NFKD', $verbose ) && $result;
 239         $result = verbosify( $c[5], $u->toNFKD( $c[5] ), 5, 'NFKD', $verbose ) && $result;
 240
 241         return $result;
 242 }
 243
 244 function testInvariant( &$u, $char, $desc, $verbose, $reportFailure = false ) {
 245         $result = verbosify( $char, $u->toNFC( $char ), 1, 'NFC', $reportFailure );
 246         $result = verbosify( $char, $u->toNFD( $char ), 1, 'NFD', $reportFailure ) && $result;
 247         $result = verbosify( $char, $u->toNFKC( $char ), 1, 'NFKC', $reportFailure ) && $result;
 248         $result = verbosify( $char, $u->toNFKD( $char ), 1, 'NFKD', $reportFailure ) && $result;
 249         $result = verbosify( $char, $u->cleanUp( $char ), 1, 'cleanUp', $reportFailure ) && $result;
 250
 251         if ( $verbose && !$result && !$reportFailure ) {
 252                 print $desc;
 253                 testInvariant( $u, $char, $desc, $verbose, true );
 254         }
 255
 256         return $result;
 257 }