languages/LanguageUtf8.php

   1 <?php
   2 /**
   3   * @package MediaWiki
   4   * @subpackage Language
   5   */
   6
   7 if( defined( "MEDIAWIKI" ) ) {
   8
   9 # This file and LanguageLatin1.php may be included from within functions, so
  10 # we need to have global statements
  11
  12 global $wgInputEncoding, $wgOutputEncoding, $wikiUpperChars, $wikiLowerChars;
  13 global $wgDBname, $wgMemc;
  14
  15 $wgInputEncoding    = "UTF-8";
  16 $wgOutputEncoding       = "UTF-8";
  17
  18 if( function_exists( 'mb_strtoupper' ) ) {
  19         mb_internal_encoding('UTF-8');
  20 } else {
  21         # Hack our own case conversion routines
  22
  23         # Loading serialized arrays is faster than parsing code :P
  24         $wikiUpperChars = $wgMemc->get( $key1 = "$wgDBname:utf8:upper" );
  25         $wikiLowerChars = $wgMemc->get( $key2 = "$wgDBname:utf8:lower" );
  26
  27         if(empty( $wikiUpperChars) || empty($wikiLowerChars )) {
  28                 require_once( "includes/Utf8Case.php" );
  29                 $wgMemc->set( $key1, $wikiUpperChars );
  30                 $wgMemc->set( $key2, $wikiLowerChars );
  31         }
  32 }
  33
  34 /**
  35  * Base stuff useful to all UTF-8 based language files
  36  * @package MediaWiki
  37  */
  38 class LanguageUtf8 extends Language {
  39
  40         # These two functions use mbstring library, if it is loaded
  41         # or compiled and character mapping arrays otherwise.
  42         # In case of language-specific character mismatch
  43         # it should be dealt with in Language classes.
  44
  45         function ucfirst( $string ) {
  46                 /**
  47                  * On pages with many links we can get called a lot.
  48                  * The multibyte uppercase functions are relatively
  49                  * slow, so check first if we can use a faster ASCII
  50                  * version instead; it saves a few milliseconds.
  51                  */
  52                 if( preg_match( '/^[\x80-\xff]/', $string ) ) {
  53                         if (function_exists('mb_strtoupper')) {
  54                                 return mb_strtoupper(mb_substr($string,0,1)).mb_substr($string,1);
  55                         } else {
  56                                 global $wikiUpperChars;
  57                                 return preg_replace (
  58                                         "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/e",
  59                                         "strtr ( \"\$1\" , \$wikiUpperChars )",
  60                                         $string );
  61                         }
  62                 }
  63                 return ucfirst( $string );
  64         }
  65
  66         function lcfirst( $string ) {
  67                 if (function_exists('mb_strtolower')) {
  68                         return mb_strtolower(mb_substr($string,0,1)).mb_substr($string,1);
  69                 } else {
  70                     global $wikiLowerChars;
  71                     return preg_replace (
  72                     "/^([A-Z]|[\\xc0-\\xff][\\x80-\\xbf]*)/e",
  73                     "strtr ( \"\$1\" , \$wikiLowerChars )",
  74                     $string );
  75                 }
  76         }
  77
  78         function stripForSearch( $string ) {
  79                 # MySQL fulltext index doesn't grok utf-8, so we
  80                 # need to fold cases and convert to hex
  81
  82                 # In Language:: it just returns lowercase, maybe
  83                 # all strtolower on stripped output or argument
  84                 # should be removed and all stripForSearch
  85                 # methods adjusted to that.
  86
  87                 wfProfileIn( "LanguageUtf8::stripForSearch" );
  88                 if( function_exists( 'mb_strtolower' ) ) {
  89                         $out = preg_replace(
  90                                 "/([\\xc0-\\xff][\\x80-\\xbf]*)/e",
  91                                 "'U8' . bin2hex( \"$1\" )",
  92                                 mb_strtolower( $string ) );
  93                 } else {
  94                         global $wikiLowerChars;
  95                         $out = preg_replace(
  96                                 "/([\\xc0-\\xff][\\x80-\\xbf]*)/e",
  97                                 "'U8' . bin2hex( strtr( \"\$1\", \$wikiLowerChars ) )",
  98                                 $string );
  99                 }
 100                 wfProfileOut( "LanguageUtf8::stripForSearch" );
 101                 return $out;
 102         }
 103
 104         function fallback8bitEncoding() {
 105                 # Windows codepage 1252 is a superset of iso 8859-1
 106                 # override this to use difference source encoding to
 107                 # translate incoming 8-bit URLs.
 108                 return "windows-1252";
 109         }
 110
 111         function checkTitleEncoding( $s ) {
 112                 global $wgInputEncoding;
 113
 114                 if( is_array( $s ) ) {
 115                         wfDebugDieBacktrace( 'Given array to checkTitleEncoding.' );
 116                 }
 117                 # Check for non-UTF-8 URLs
 118                 $ishigh = preg_match( '/[\x80-\xff]/', $s);
 119                 if(!$ishigh) return $s;
 120
 121                 $isutf8 = preg_match( '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
 122                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})+$/', $s );
 123                 if( $isutf8 ) return $s;
 124
 125                 return $this->iconv( $this->fallback8bitEncoding(), "utf-8", $s );
 126         }
 127
 128         function firstChar( $s ) {
 129                 preg_match( '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
 130                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/', $s, $matches);
 131
 132                 return isset( $matches[1] ) ? $matches[1] : "";
 133         }
 134
 135         # Crop a string from the beginning or end to a certain number of bytes.
 136         # (Bytes are used because our storage has limited byte lengths for some
 137         # columns in the database.) Multibyte charsets will need to make sure that
 138         # only whole characters are included!
 139         #
 140         # $length does not include the optional ellipsis.
 141         # If $length is negative, snip from the beginning
 142         function truncate( $string, $length, $ellipsis = "" ) {
 143                 if( $length == 0 ) {
 144                         return $ellipsis;
 145                 }
 146                 if ( strlen( $string ) <= abs( $length ) ) {
 147                         return $string;
 148                 }
 149                 if( $length > 0 ) {
 150                         $string = substr( $string, 0, $length );
 151                         $char = ord( $string[strlen( $string ) - 1] );
 152                         if ($char >= 0xc0) {
 153                                 # We got the first byte only of a multibyte char; remove it.
 154                                 $string = substr( $string, 0, -1 );
 155                         } elseif( $char >= 0x80 &&
 156                                   preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
 157                                               '[\xf0-\xf7][\x80-\xbf]{1,2})$/', $string, $m ) ) {
 158                             # We chopped in the middle of a character; remove it
 159                                 $string = $m[1];
 160                         }
 161                         return $string . $ellipsis;
 162                 } else {
 163                         $string = substr( $string, $length );
 164                         $char = ord( $string[0] );
 165                         if( $char >= 0x80 && $char < 0xc0 ) {
 166                                 # We chopped in the middle of a character; remove the whole thing
 167                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
 168                         }
 169                         return $ellipsis . $string;
 170                 }
 171         }
 172 }
 173
 174 } # ifdef MEDIAWIKI
 175
 176 ?>