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