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 # 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  18 # http://www.gnu.org/copyleft/gpl.html
  19
  20 # Implements the conformance test at:
  21 # http://www.unicode.org/Public/UNIDATA/NormalizationTest.txt
  22
  23 $verbose = true;
  24 #define( 'PRETTY_UTF8', true );
  25
  26 if( defined( 'PRETTY_UTF8' ) ) {
  27         function pretty( $string ) {
  28                 return preg_replace( '/([\x00-\xff])/e',
  29                         'sprintf("%02X", ord("$1"))',
  30                         $string );
  31         }
  32 } else {
  33         function pretty( $string ) {
  34                 return trim( preg_replace( '/(.)/use',
  35                         'sprintf("%04X ", utf8ToCodepoint("$1"))',
  36                         $string ) );
  37         }
  38 }
  39
  40 require_once 'UtfNormalUtil.php';
  41 require_once 'UtfNormal.php';
  42
  43 if( php_sapi_name() != 'cli' ) {
  44         die( "Run me from the command line please.\n" );
  45 }
  46
  47 $in = fopen("NormalizationTest.txt", "rt");
  48 if( !$in ) {
  49         print "Couldn't open NormalizationTest.txt -- can't run tests.\n";
  50         print "If necessary, manually download this file. It can be obtained at\n";
  51         print "http://www.unicode.org/Public/UNIDATA/NormalizationTest.txt";
  52         exit(-1);
  53 }
  54
  55 $normalizer = new UtfNormal;
  56
  57 $total = 0;
  58 $success = 0;
  59 $failure = 0;
  60 $ok = true;
  61 $testedChars = array();
  62 while( false !== ( $line = fgets( $in ) ) ) {
  63         list( $data, $comment ) = explode( '#', $line );
  64         if( $data === '' ) continue;
  65         if( preg_match( '/@Part([\d])/', $data, $matches ) ) {
  66                 if( $matches[1] > 0 ) {
  67                         $ok = reportResults( $total, $success, $failure ) && $ok;
  68                 }
  69                 print "Part {$matches[1]}: $comment";
  70                 continue;
  71         }
  72
  73         $columns = array_map( "hexSequenceToUtf8", explode( ";", $data ) );
  74         array_unshift( $columns, '' );
  75
  76         $testedChars[$columns[1]] = true;
  77         $total++;
  78         if( testNormals( $normalizer, $columns, $comment ) ) {
  79                 $success++;
  80         } else {
  81                 $failure++;
  82                 # print "FAILED: $comment";
  83         }
  84         if( $total % 100 == 0 ) print "$total ";
  85 }
  86 fclose( $in );
  87
  88 $ok = reportResults( $total, $success, $failure ) && $ok;
  89
  90 $in = fopen("UnicodeData.txt", "rt" );
  91 if( !$in ) {
  92         print "Can't open UnicodeData.txt for reading.\n";
  93         print "If necessary, fetch this file from the internet:\n";
  94         print "http://www.unicode.org/Public/UNIDATA/UnicodeData.txt\n";
  95         exit(-1);
  96 }
  97 print "Now testing invariants...\n";
  98 while( false !== ($line = fgets( $in ) ) ) {
  99         $cols = explode( ';', $line );
 100         $char = codepointToUtf8( hexdec( $cols[0] ) );
 101         $desc = $cols[0] . ": " . $cols[1];
 102         if( empty( $testedChars[$char] ) ) {
 103                 $total++;
 104                 if( testInvariant( $normalizer, $char, $desc ) ) {
 105                         $success++;
 106                 } else {
 107                         $failure++;
 108                 }
 109                 if( $total % 100 == 0 ) print "$total ";
 110         }
 111 }
 112 fclose( $in );
 113
 114 $ok = reportResults( $total, $success, $failure ) && $ok;
 115
 116 if( $ok ) {
 117         print "TEST SUCCEEDED!\n";
 118         exit(0);
 119 } else {
 120         print "TEST FAILED!\n";
 121         exit(-1);
 122 }
 123
 124 ## ------
 125
 126 function reportResults( &$total, &$success, &$failure ) {
 127         $percSucc = IntVal( $success * 100 / $total );
 128         $percFail = IntVal( $failure * 100 / $total );
 129         print "\n";
 130         print "$success tests successful ($percSucc%)\n";
 131         print "$failure tests failed ($percFail%)\n\n";
 132         $ok = ($success > 0 && $failure == 0);
 133         $total = 0;
 134         $success = 0;
 135         $failure = 0;
 136         return $ok;
 137 }
 138
 139 function testNormals( &$u, $c, $comment, $reportFailure = false ) {
 140         $result = testNFC( $u, $c, $comment, $reportFailure );
 141         $result = testNFD( $u, $c, $comment, $reportFailure ) && $result;
 142         $result = testNFKC( $u, $c, $comment, $reportFailure ) && $result;
 143         $result = testNFKD( $u, $c, $comment, $reportFailure ) && $result;
 144
 145         global $verbose;
 146         if( $verbose && !$result && !$reportFailure ) {
 147                 print $comment;
 148                 testNormals( $u, $c, $comment, true );
 149         }
 150         return $result;
 151 }
 152
 153 function verbosify( $a, $b, $col, $form, $verbose ) {
 154         #$result = ($a === $b);
 155         $result = (strcmp( $a, $b ) == 0);
 156         if( $verbose ) {
 157                 $aa = pretty( $a );
 158                 $bb = pretty( $b );
 159                 $ok = $result ? "succeed" : " failed";
 160                 $eq = $result ? "==" : "!=";
 161                 print "  $ok $form c$col '$aa' $eq '$bb'\n";
 162         }
 163         return $result;
 164 }
 165
 166 function testNFC( &$u, $c, $comment, $verbose ) {
 167         $result = verbosify( $c[2], $u->toNFC( $c[1] ), 1, 'NFC', $verbose );
 168         $result = verbosify( $c[2], $u->toNFC( $c[2] ), 2, 'NFC', $verbose ) && $result;
 169         $result = verbosify( $c[2], $u->toNFC( $c[3] ), 3, 'NFC', $verbose ) && $result;
 170         $result = verbosify( $c[4], $u->toNFC( $c[4] ), 4, 'NFC', $verbose ) && $result;
 171         $result = verbosify( $c[4], $u->toNFC( $c[5] ), 5, 'NFC', $verbose ) && $result;
 172         return $result;
 173 }
 174
 175 function testNFD( &$u, $c, $comment, $verbose ) {
 176         $result = verbosify( $c[3], $u->toNFD( $c[1] ), 1, 'NFD', $verbose );
 177         $result = verbosify( $c[3], $u->toNFD( $c[2] ), 2, 'NFD', $verbose ) && $result;
 178         $result = verbosify( $c[3], $u->toNFD( $c[3] ), 3, 'NFD', $verbose ) && $result;
 179         $result = verbosify( $c[5], $u->toNFD( $c[4] ), 4, 'NFD', $verbose ) && $result;
 180         $result = verbosify( $c[5], $u->toNFD( $c[5] ), 5, 'NFD', $verbose ) && $result;
 181         return $result;
 182 }
 183
 184 function testNFKC( &$u, $c, $comment, $verbose ) {
 185         $result = verbosify( $c[4], $u->toNFKC( $c[1] ), 1, 'NFKC', $verbose );
 186         $result = verbosify( $c[4], $u->toNFKC( $c[2] ), 2, 'NFKC', $verbose ) && $result;
 187         $result = verbosify( $c[4], $u->toNFKC( $c[3] ), 3, 'NFKC', $verbose ) && $result;
 188         $result = verbosify( $c[4], $u->toNFKC( $c[4] ), 4, 'NFKC', $verbose ) && $result;
 189         $result = verbosify( $c[4], $u->toNFKC( $c[5] ), 5, 'NFKC', $verbose ) && $result;
 190         return $result;
 191 }
 192
 193 function testNFKD( &$u, $c, $comment, $verbose ) {
 194         $result = verbosify( $c[5], $u->toNFKD( $c[1] ), 1, 'NFKD', $verbose );
 195         $result = verbosify( $c[5], $u->toNFKD( $c[2] ), 2, 'NFKD', $verbose ) && $result;
 196         $result = verbosify( $c[5], $u->toNFKD( $c[3] ), 3, 'NFKD', $verbose ) && $result;
 197         $result = verbosify( $c[5], $u->toNFKD( $c[4] ), 4, 'NFKD', $verbose ) && $result;
 198         $result = verbosify( $c[5], $u->toNFKD( $c[5] ), 5, 'NFKD', $verbose ) && $result;
 199         return $result;
 200 }
 201
 202 function testInvariant( &$u, $char, $desc, $reportFailure = false ) {
 203         $result = verbosify( $char, $u->toNFC( $char ), 1, 'NFC', $reportFailure );
 204         $result = verbosify( $char, $u->toNFD( $char ), 1, 'NFD', $reportFailure ) && $result;
 205         $result = verbosify( $char, $u->toNFKC( $char ), 1, 'NFKC', $reportFailure ) && $result;
 206         $result = verbosify( $char, $u->toNFKD( $char ), 1, 'NFKD', $reportFailure ) && $result;
 207         global $verbose;
 208         if( $verbose && !$result && !$reportFailure ) {
 209                 print $desc;
 210                 testInvariant( $u, $char, $desc, true );
 211         }
 212         return $result;
 213 }
 214
 215 ?>