includes/normal/UtfNormalTest.php

   1 <?php
   2 # Copyright (C) 2004 Brion Vibber <brion@pobox.com>
   3 # http://www.mediawiki.org/
   4 #
   5 # This program is free software; you can redistribute it and/or modify
   6 # it under the terms of the GNU General Public License as published by
   7 # the Free Software Foundation; either version 2 of the License, or
   8 # (at your option) any later version.
   9 #
  10 # This program is distributed in the hope that it will be useful,
  11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13 # GNU General Public License for more details.
  14 #
  15 # You should have received a copy of the GNU General Public License along
  16 # with this program; if not, write to the Free Software Foundation, Inc.,
  17 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18 # http://www.gnu.org/copyleft/gpl.html
  19
  20 /**
  21  * Implements the conformance test at:
  22  * http://www.unicode.org/Public/UNIDATA/NormalizationTest.txt
  23  * @ingroup UtfNormal
  24  */
  25
  26 /** */
  27 $verbose = true;
  28 #define( 'PRETTY_UTF8', true );
  29
  30 if( defined( 'PRETTY_UTF8' ) ) {
  31         function pretty( $string ) {
  32                 return preg_replace( '/([\x00-\xff])/e',
  33                         'sprintf("%02X", ord("$1"))',
  34                         $string );
  35         }
  36 } else {
  37         /**
  38          * @ignore
  39          */
  40         function pretty( $string ) {
  41                 return trim( preg_replace( '/(.)/use',
  42                         'sprintf("%04X ", utf8ToCodepoint("$1"))',
  43                         $string ) );
  44         }
  45 }
  46
  47 if( isset( $_SERVER['argv'] ) && in_array( '--icu', $_SERVER['argv'] ) ) {
  48         dl( 'php_utfnormal.so' );
  49 }
  50
  51 require_once 'UtfNormalUtil.php';
  52 require_once 'UtfNormal.php';
  53
  54 if( php_sapi_name() != 'cli' ) {
  55         die( "Run me from the command line please.\n" );
  56 }
  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 while( false !== ( $line = fgets( $in ) ) ) {
  74         list( $data, $comment ) = explode( '#', $line );
  75         if( $data === '' ) continue;
  76         $matches = array();
  77         if( preg_match( '/@Part([\d])/', $data, $matches ) ) {
  78                 if( $matches[1] > 0 ) {
  79                         $ok = reportResults( $total, $success, $failure ) && $ok;
  80                 }
  81                 print "Part {$matches[1]}: $comment";
  82                 continue;
  83         }
  84
  85         $columns = array_map( "hexSequenceToUtf8", explode( ";", $data ) );
  86         array_unshift( $columns, '' );
  87
  88         $testedChars[$columns[1]] = true;
  89         $total++;
  90         if( testNormals( $normalizer, $columns, $comment ) ) {
  91                 $success++;
  92         } else {
  93                 $failure++;
  94                 # print "FAILED: $comment";
  95         }
  96         if( $total % 100 == 0 ) print "$total ";
  97 }
  98 fclose( $in );
  99
 100 $ok = reportResults( $total, $success, $failure ) && $ok;
 101
 102 $in = fopen("UnicodeData.txt", "rt" );
 103 if( !$in ) {
 104         print "Can't open UnicodeData.txt for reading.\n";
 105         print "If necessary, fetch this file from the internet:\n";
 106         print "http://www.unicode.org/Public/UNIDATA/UnicodeData.txt\n";
 107         exit(-1);
 108 }
 109 print "Now testing invariants...\n";
 110 while( false !== ($line = fgets( $in ) ) ) {
 111         $cols = explode( ';', $line );
 112         $char = codepointToUtf8( hexdec( $cols[0] ) );
 113         $desc = $cols[0] . ": " . $cols[1];
 114         if( $char < "\x20" || $char >= UTF8_SURROGATE_FIRST && $char <= UTF8_SURROGATE_LAST ) {
 115                 # Can't check NULL with the ICU plugin, as null bytes fail in C land.
 116                 # Skip other control characters, as we strip them for XML safety.
 117                 # Surrogates are illegal on their own or in UTF-8, ignore.
 118                 continue;
 119         }
 120         if( empty( $testedChars[$char] ) ) {
 121                 $total++;
 122                 if( testInvariant( $normalizer, $char, $desc ) ) {
 123                         $success++;
 124                 } else {
 125                         $failure++;
 126                 }
 127                 if( $total % 100 == 0 ) print "$total ";
 128         }
 129 }
 130 fclose( $in );
 131
 132 $ok = reportResults( $total, $success, $failure ) && $ok;
 133
 134 if( $ok ) {
 135         print "TEST SUCCEEDED!\n";
 136         exit(0);
 137 } else {
 138         print "TEST FAILED!\n";
 139         exit(-1);
 140 }
 141
 142 ## ------
 143
 144 function reportResults( &$total, &$success, &$failure ) {
 145         $percSucc = intval( $success * 100 / $total );
 146         $percFail = intval( $failure * 100 / $total );
 147         print "\n";
 148         print "$success tests successful ($percSucc%)\n";
 149         print "$failure tests failed ($percFail%)\n\n";
 150         $ok = ($success > 0 && $failure == 0);
 151         $total = 0;
 152         $success = 0;
 153         $failure = 0;
 154         return $ok;
 155 }
 156
 157 function testNormals( &$u, $c, $comment, $reportFailure = false ) {
 158         $result = testNFC( $u, $c, $comment, $reportFailure );
 159         $result = testNFD( $u, $c, $comment, $reportFailure ) && $result;
 160         $result = testNFKC( $u, $c, $comment, $reportFailure ) && $result;
 161         $result = testNFKD( $u, $c, $comment, $reportFailure ) && $result;
 162         $result = testCleanUp( $u, $c, $comment, $reportFailure ) && $result;
 163
 164         global $verbose;
 165         if( $verbose && !$result && !$reportFailure ) {
 166                 print $comment;
 167                 testNormals( $u, $c, $comment, 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, $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         global $verbose;
 242         if( $verbose && !$result && !$reportFailure ) {
 243                 print $desc;
 244                 testInvariant( $u, $char, $desc, true );
 245         }
 246         return $result;
 247 }