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 $verbose = true;
  29 #define( 'PRETTY_UTF8', true );
  30
  31 if( defined( 'PRETTY_UTF8' ) ) {
  32         function pretty( $string ) {
  33                 return preg_replace( '/([\x00-\xff])/e',
  34                         'sprintf("%02X", ord("$1"))',
  35                         $string );
  36         }
  37 } else {
  38         /**
  39          * @ignore
  40          * @return string
  41          */
  42         function pretty( $string ) {
  43                 return trim( preg_replace( '/(.)/use',
  44                         'sprintf("%04X ", utf8ToCodepoint("$1"))',
  45                         $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 if( php_sapi_name() != 'cli' ) {
  58         die( "Run me from the command line please.\n" );
  59 }
  60
  61 $in = fopen("NormalizationTest.txt", "rt");
  62 if( !$in ) {
  63         print "Couldn't open NormalizationTest.txt -- can't run tests.\n";
  64         print "If necessary, manually download this file. It can be obtained at\n";
  65         print "http://www.unicode.org/Public/UNIDATA/NormalizationTest.txt";
  66         exit(-1);
  67 }
  68
  69 $normalizer = new UtfNormal;
  70
  71 $total = 0;
  72 $success = 0;
  73 $failure = 0;
  74 $ok = true;
  75 $testedChars = array();
  76 while( false !== ( $line = fgets( $in ) ) ) {
  77         list( $data, $comment ) = explode( '#', $line );
  78         if( $data === '' ) continue;
  79         $matches = array();
  80         if( preg_match( '/@Part([\d])/', $data, $matches ) ) {
  81                 if( $matches[1] > 0 ) {
  82                         $ok = reportResults( $total, $success, $failure ) && $ok;
  83                 }
  84                 print "Part {$matches[1]}: $comment";
  85                 continue;
  86         }
  87
  88         $columns = array_map( "hexSequenceToUtf8", explode( ";", $data ) );
  89         array_unshift( $columns, '' );
  90
  91         $testedChars[$columns[1]] = true;
  92         $total++;
  93         if( testNormals( $normalizer, $columns, $comment, $verbose ) ) {
  94                 $success++;
  95         } else {
  96                 $failure++;
  97                 # print "FAILED: $comment";
  98         }
  99         if( $total % 100 == 0 ) print "$total ";
 100 }
 101 fclose( $in );
 102
 103 $ok = reportResults( $total, $success, $failure ) && $ok;
 104
 105 $in = fopen("UnicodeData.txt", "rt" );
 106 if( !$in ) {
 107         print "Can't open UnicodeData.txt for reading.\n";
 108         print "If necessary, fetch this file from the internet:\n";
 109         print "http://www.unicode.org/Public/UNIDATA/UnicodeData.txt\n";
 110         exit(-1);
 111 }
 112 print "Now testing invariants...\n";
 113 while( false !== ($line = fgets( $in ) ) ) {
 114         $cols = explode( ';', $line );
 115         $char = codepointToUtf8( hexdec( $cols[0] ) );
 116         $desc = $cols[0] . ": " . $cols[1];
 117         if( $char < "\x20" || $char >= UTF8_SURROGATE_FIRST && $char <= UTF8_SURROGATE_LAST ) {
 118                 # Can't check NULL with the ICU plugin, as null bytes fail in C land.
 119                 # Skip other control characters, as we strip them for XML safety.
 120                 # Surrogates are illegal on their own or in UTF-8, ignore.
 121                 continue;
 122         }
 123         if( empty( $testedChars[$char] ) ) {
 124                 $total++;
 125                 if( testInvariant( $normalizer, $char, $desc, $verbose ) ) {
 126                         $success++;
 127                 } else {
 128                         $failure++;
 129                 }
 130                 if( $total % 100 == 0 ) print "$total ";
 131         }
 132 }
 133 fclose( $in );
 134
 135 $ok = reportResults( $total, $success, $failure ) && $ok;
 136
 137 if( $ok ) {
 138         print "TEST SUCCEEDED!\n";
 139         exit(0);
 140 } else {
 141         print "TEST FAILED!\n";
 142         exit(-1);
 143 }
 144
 145 ## ------
 146
 147 function reportResults( &$total, &$success, &$failure ) {
 148         $percSucc = intval( $success * 100 / $total );
 149         $percFail = intval( $failure * 100 / $total );
 150         print "\n";
 151         print "$success tests successful ($percSucc%)\n";
 152         print "$failure tests failed ($percFail%)\n\n";
 153         $ok = ($success > 0 && $failure == 0);
 154         $total = 0;
 155         $success = 0;
 156         $failure = 0;
 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         return $result;
 172 }
 173
 174 function verbosify( $a, $b, $col, $form, $verbose ) {
 175         #$result = ($a === $b);
 176         $result = (strcmp( $a, $b ) == 0);
 177         if( $verbose ) {
 178                 $aa = pretty( $a );
 179                 $bb = pretty( $b );
 180                 $ok = $result ? "succeed" : " failed";
 181                 $eq = $result ? "==" : "!=";
 182                 print "  $ok $form c$col '$aa' $eq '$bb'\n";
 183         }
 184         return $result;
 185 }
 186
 187 function testNFC( &$u, $c, $comment, $verbose ) {
 188         $result = verbosify( $c[2], $u->toNFC( $c[1] ), 1, 'NFC', $verbose );
 189         $result = verbosify( $c[2], $u->toNFC( $c[2] ), 2, 'NFC', $verbose ) && $result;
 190         $result = verbosify( $c[2], $u->toNFC( $c[3] ), 3, 'NFC', $verbose ) && $result;
 191         $result = verbosify( $c[4], $u->toNFC( $c[4] ), 4, 'NFC', $verbose ) && $result;
 192         $result = verbosify( $c[4], $u->toNFC( $c[5] ), 5, 'NFC', $verbose ) && $result;
 193         return $result;
 194 }
 195
 196 function testCleanUp( &$u, $c, $comment, $verbose ) {
 197         $x = $c[1];
 198         $result = verbosify( $c[2], $u->cleanUp( $x ), 1, 'cleanUp', $verbose );
 199         $x = $c[2];
 200         $result = verbosify( $c[2], $u->cleanUp( $x ), 2, 'cleanUp', $verbose ) && $result;
 201         $x = $c[3];
 202         $result = verbosify( $c[2], $u->cleanUp( $x ), 3, 'cleanUp', $verbose ) && $result;
 203         $x = $c[4];
 204         $result = verbosify( $c[4], $u->cleanUp( $x ), 4, 'cleanUp', $verbose ) && $result;
 205         $x = $c[5];
 206         $result = verbosify( $c[4], $u->cleanUp( $x ), 5, 'cleanUp', $verbose ) && $result;
 207         return $result;
 208 }
 209
 210 function testNFD( &$u, $c, $comment, $verbose ) {
 211         $result = verbosify( $c[3], $u->toNFD( $c[1] ), 1, 'NFD', $verbose );
 212         $result = verbosify( $c[3], $u->toNFD( $c[2] ), 2, 'NFD', $verbose ) && $result;
 213         $result = verbosify( $c[3], $u->toNFD( $c[3] ), 3, 'NFD', $verbose ) && $result;
 214         $result = verbosify( $c[5], $u->toNFD( $c[4] ), 4, 'NFD', $verbose ) && $result;
 215         $result = verbosify( $c[5], $u->toNFD( $c[5] ), 5, 'NFD', $verbose ) && $result;
 216         return $result;
 217 }
 218
 219 function testNFKC( &$u, $c, $comment, $verbose ) {
 220         $result = verbosify( $c[4], $u->toNFKC( $c[1] ), 1, 'NFKC', $verbose );
 221         $result = verbosify( $c[4], $u->toNFKC( $c[2] ), 2, 'NFKC', $verbose ) && $result;
 222         $result = verbosify( $c[4], $u->toNFKC( $c[3] ), 3, 'NFKC', $verbose ) && $result;
 223         $result = verbosify( $c[4], $u->toNFKC( $c[4] ), 4, 'NFKC', $verbose ) && $result;
 224         $result = verbosify( $c[4], $u->toNFKC( $c[5] ), 5, 'NFKC', $verbose ) && $result;
 225         return $result;
 226 }
 227
 228 function testNFKD( &$u, $c, $comment, $verbose ) {
 229         $result = verbosify( $c[5], $u->toNFKD( $c[1] ), 1, 'NFKD', $verbose );
 230         $result = verbosify( $c[5], $u->toNFKD( $c[2] ), 2, 'NFKD', $verbose ) && $result;
 231         $result = verbosify( $c[5], $u->toNFKD( $c[3] ), 3, 'NFKD', $verbose ) && $result;
 232         $result = verbosify( $c[5], $u->toNFKD( $c[4] ), 4, 'NFKD', $verbose ) && $result;
 233         $result = verbosify( $c[5], $u->toNFKD( $c[5] ), 5, 'NFKD', $verbose ) && $result;
 234         return $result;
 235 }
 236
 237 function testInvariant( &$u, $char, $desc, $verbose, $reportFailure = false ) {
 238         $result = verbosify( $char, $u->toNFC( $char ), 1, 'NFC', $reportFailure );
 239         $result = verbosify( $char, $u->toNFD( $char ), 1, 'NFD', $reportFailure ) && $result;
 240         $result = verbosify( $char, $u->toNFKC( $char ), 1, 'NFKC', $reportFailure ) && $result;
 241         $result = verbosify( $char, $u->toNFKD( $char ), 1, 'NFKD', $reportFailure ) && $result;
 242         $result = verbosify( $char, $u->cleanUp( $char ), 1, 'cleanUp', $reportFailure ) && $result;
 243
 244         if( $verbose && !$result && !$reportFailure ) {
 245                 print $desc;
 246                 testInvariant( $u, $char, $desc, $verbose, true );
 247         }
 248         return $result;
 249 }