languages/LanguageUtf8.php

   1 <?php
   2
   3 $wgInputEncoding    = "utf-8";
   4 $wgOutputEncoding       = "utf-8";
   5
   6 $wikiUpperChars = $wgMemc->get( $key1 = "$wgDBname:utf8:upper" );
   7 $wikiLowerChars = $wgMemc->get( $key2 = "$wgDBname:utf8:lower" );
   8
   9 if(empty( $wikiUpperChars) || empty($wikiLowerChars )) {
  10         require_once( "Utf8Case.php" );
  11         $wgMemc->set( $key1, $wikiUpperChars );
  12         $wgMemc->set( $key2, $wikiLowerChars );
  13 }
  14
  15 # Base stuff useful to all UTF-8 based language files
  16 class LanguageUtf8 extends Language {
  17
  18         function ucfirst( $string ) {
  19                 # For most languages, this is a wrapper for ucfirst()
  20                 # But that doesn't work right in a UTF-8 locale
  21                 global $wikiUpperChars, $wikiLowerChars;
  22                 return preg_replace (
  23                 "/^([\\x00-\\x7f]|[\\xc0-\\xff][\\x80-\\xbf]*)/e",
  24                 "strtr ( \"\$1\" , \$wikiUpperChars )",
  25                 $string );
  26         }
  27
  28         function lcfirst( $string ) {
  29                 global $wikiUpperChars, $wikiLowerChars;
  30                 return preg_replace (
  31                 "/^([\\x00-\\x7f]|[\\xc0-\\xff][\\x80-\\xbf]*)/e",
  32                 "strtr ( \"\$1\" , \$wikiLowerChars )",
  33                 $string );
  34         }
  35
  36         function stripForSearch( $string ) {
  37                 # MySQL fulltext index doesn't grok utf-8, so we
  38                 # need to fold cases and convert to hex
  39                 global $wikiLowerChars;
  40                 return preg_replace(
  41                   "/([\\xc0-\\xff][\\x80-\\xbf]*)/e",
  42                   "'U8' . bin2hex( strtr( \"\$1\", \$wikiLowerChars ) )",
  43                   $string );
  44         }
  45
  46         function fallback8bitEncoding() {
  47                 # Windows codepage 1252 is a superset of iso 8859-1
  48                 # override this to use difference source encoding to
  49                 # translate incoming 8-bit URLs.
  50                 return "windows-1252";
  51         }
  52
  53         function checkTitleEncoding( $s ) {
  54                 global $wgInputEncoding;
  55
  56                 # Check for non-UTF-8 URLs
  57                 $ishigh = preg_match( '/[\x80-\xff]/', $s);
  58                 if(!$ishigh) return $s;
  59
  60                 $isutf8 = preg_match( '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
  61                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})+$/', $s );
  62                 if( $isutf8 ) return $s;
  63
  64                 return $this->iconv( $this->fallback8bitEncoding(), "utf-8", $s );
  65         }
  66 }
  67
  68 ?>