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          */
  41         function pretty( $string ) {
  42                 return trim( preg_replace( '/(.)/use',
  43                         'sprintf("%04X ", utf8ToCodepoint("$1"))',
  44                         $string ) );
  45         }
  46 }
  47
  48 if( isset( $_SERVER['argv'] ) && in_array( '--icu', $_SERVER['argv'] ) ) {
  49         dl( 'php_utfnormal.so' );
  50 }
  51
  52 require_once 'UtfNormalUtil.php';
  53 require_once 'UtfNormal.php';
  54
  55 if( php_sapi_name() != 'cli' ) {
  56         die( "Run me from the command line please.\n" );
  57 }
  58
  59 $in = fopen("NormalizationTest.txt", "rt");
  60 if( !$in ) {
  61         print "Couldn't open NormalizationTest.txt -- can't run tests.\n";
  62         print "If necessary, manually download this file. It can be obtained at\n";
  63         print "http://www.unicode.org/Public/UNIDATA/NormalizationTest.txt";
  64         exit(-1);
  65 }
  66
  67 $normalizer = new UtfNormal;
  68
  69 $total = 0;
  70 $success = 0;
  71 $failure = 0;
  72 $ok = true;
  73 $testedChars = array();
  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 while( false !== ($line = fgets( $in ) ) ) {
 112         $cols = explode( ';', $line );
 113         $char = codepointToUtf8( hexdec( $cols[0] ) );
 114         $desc = $cols[0] . ": " . $cols[1];
 115         if( $char < "\x20" || $char >= UTF8_SURROGATE_FIRST && $char <= UTF8_SURROGATE_LAST ) {
 116                 # Can't check NULL with the ICU plugin, as null bytes fail in C land.
 117                 # Skip other control characters, as we strip them for XML safety.
 118                 # Surrogates are illegal on their own or in UTF-8, ignore.
 119                 continue;
 120         }
 121         if( empty( $testedChars[$char] ) ) {
 122                 $total++;
 123                 if( testInvariant( $normalizer, $char, $desc, $verbose ) ) {
 124                         $success++;
 125                 } else {
 126                         $failure++;
 127                 }
 128                 if( $total % 100 == 0 ) print "$total ";
 129         }
 130 }
 131 fclose( $in );
 132
 133 $ok = reportResults( $total, $success, $failure ) && $ok;
 134
 135 if( $ok ) {
 136         print "TEST SUCCEEDED!\n";
 137         exit(0);
 138 } else {
 139         print "TEST FAILED!\n";
 140         exit(-1);
 141 }
 142
 143 ## ------
 144
 145 function reportResults( &$total, &$success, &$failure ) {
 146         $percSucc = intval( $success * 100 / $total );
 147         $percFail = intval( $failure * 100 / $total );
 148         print "\n";
 149         print "$success tests successful ($percSucc%)\n";
 150         print "$failure tests failed ($percFail%)\n\n";
 151         $ok = ($success > 0 && $failure == 0);
 152         $total = 0;
 153         $success = 0;
 154         $failure = 0;
 155         return $ok;
 156 }
 157
 158 function testNormals( &$u, $c, $comment, $verbose, $reportFailure = false ) {
 159         $result = testNFC( $u, $c, $comment, $reportFailure );
 160         $result = testNFD( $u, $c, $comment, $reportFailure ) && $result;
 161         $result = testNFKC( $u, $c, $comment, $reportFailure ) && $result;
 162         $result = testNFKD( $u, $c, $comment, $reportFailure ) && $result;
 163         $result = testCleanUp( $u, $c, $comment, $reportFailure ) && $result;
 164
 165         if( $verbose && !$result && !$reportFailure ) {
 166                 print $comment;
 167                 testNormals( $u, $c, $comment, $verbose, true );
 168         }
 169         return $result;
 170 }
 171
 172 function verbosify( $a, $b, $col, $form, $verbose ) {
 173         #$result = ($a === $b);
 174         $result = (strcmp( $a, $b ) == 0);
 175         if( $verbose ) {
 176                 $aa = pretty( $a );
 177                 $bb = pretty( $b );
 178                 $ok = $result ? "succeed" : " failed";
 179                 $eq = $result ? "==" : "!=";
 180                 print "  $ok $form c$col '$aa' $eq '$bb'\n";
 181         }
 182         return $result;
 183 }
 184
 185 function testNFC( &$u, $c, $comment, $verbose ) {
 186         $result = verbosify( $c[2], $u->toNFC( $c[1] ), 1, 'NFC', $verbose );
 187         $result = verbosify( $c[2], $u->toNFC( $c[2] ), 2, 'NFC', $verbose ) && $result;
 188         $result = verbosify( $c[2], $u->toNFC( $c[3] ), 3, 'NFC', $verbose ) && $result;
 189         $result = verbosify( $c[4], $u->toNFC( $c[4] ), 4, 'NFC', $verbose ) && $result;
 190         $result = verbosify( $c[4], $u->toNFC( $c[5] ), 5, 'NFC', $verbose ) && $result;
 191         return $result;
 192 }
 193
 194 function testCleanUp( &$u, $c, $comment, $verbose ) {
 195         $x = $c[1];
 196         $result = verbosify( $c[2], $u->cleanUp( $x ), 1, 'cleanUp', $verbose );
 197         $x = $c[2];
 198         $result = verbosify( $c[2], $u->cleanUp( $x ), 2, 'cleanUp', $verbose ) && $result;
 199         $x = $c[3];
 200         $result = verbosify( $c[2], $u->cleanUp( $x ), 3, 'cleanUp', $verbose ) && $result;
 201         $x = $c[4];
 202         $result = verbosify( $c[4], $u->cleanUp( $x ), 4, 'cleanUp', $verbose ) && $result;
 203         $x = $c[5];
 204         $result = verbosify( $c[4], $u->cleanUp( $x ), 5, 'cleanUp', $verbose ) && $result;
 205         return $result;
 206 }
 207
 208 function testNFD( &$u, $c, $comment, $verbose ) {
 209         $result = verbosify( $c[3], $u->toNFD( $c[1] ), 1, 'NFD', $verbose );
 210         $result = verbosify( $c[3], $u->toNFD( $c[2] ), 2, 'NFD', $verbose ) && $result;
 211         $result = verbosify( $c[3], $u->toNFD( $c[3] ), 3, 'NFD', $verbose ) && $result;
 212         $result = verbosify( $c[5], $u->toNFD( $c[4] ), 4, 'NFD', $verbose ) && $result;
 213         $result = verbosify( $c[5], $u->toNFD( $c[5] ), 5, 'NFD', $verbose ) && $result;
 214         return $result;
 215 }
 216
 217 function testNFKC( &$u, $c, $comment, $verbose ) {
 218         $result = verbosify( $c[4], $u->toNFKC( $c[1] ), 1, 'NFKC', $verbose );
 219         $result = verbosify( $c[4], $u->toNFKC( $c[2] ), 2, 'NFKC', $verbose ) && $result;
 220         $result = verbosify( $c[4], $u->toNFKC( $c[3] ), 3, 'NFKC', $verbose ) && $result;
 221         $result = verbosify( $c[4], $u->toNFKC( $c[4] ), 4, 'NFKC', $verbose ) && $result;
 222         $result = verbosify( $c[4], $u->toNFKC( $c[5] ), 5, 'NFKC', $verbose ) && $result;
 223         return $result;
 224 }
 225
 226 function testNFKD( &$u, $c, $comment, $verbose ) {
 227         $result = verbosify( $c[5], $u->toNFKD( $c[1] ), 1, 'NFKD', $verbose );
 228         $result = verbosify( $c[5], $u->toNFKD( $c[2] ), 2, 'NFKD', $verbose ) && $result;
 229         $result = verbosify( $c[5], $u->toNFKD( $c[3] ), 3, 'NFKD', $verbose ) && $result;
 230         $result = verbosify( $c[5], $u->toNFKD( $c[4] ), 4, 'NFKD', $verbose ) && $result;
 231         $result = verbosify( $c[5], $u->toNFKD( $c[5] ), 5, 'NFKD', $verbose ) && $result;
 232         return $result;
 233 }
 234
 235 function testInvariant( &$u, $char, $desc, $verbose, $reportFailure = false ) {
 236         $result = verbosify( $char, $u->toNFC( $char ), 1, 'NFC', $reportFailure );
 237         $result = verbosify( $char, $u->toNFD( $char ), 1, 'NFD', $reportFailure ) && $result;
 238         $result = verbosify( $char, $u->toNFKC( $char ), 1, 'NFKC', $reportFailure ) && $result;
 239         $result = verbosify( $char, $u->toNFKD( $char ), 1, 'NFKD', $reportFailure ) && $result;
 240         $result = verbosify( $char, $u->cleanUp( $char ), 1, 'cleanUp', $reportFailure ) && $result;
 241
 242         if( $verbose && !$result && !$reportFailure ) {
 243                 print $desc;
 244                 testInvariant( $u, $char, $desc, $verbose, true );
 245         }
 246         return $result;
 247 }