functions/i18n.php

   1 <?
   2
   3 /**
   4  ** i18n.php
   5  **
   6  ** This file contains variuos functions that are needed to do
   7  ** internationalization of SquirrelMail.
   8  **
   9  ** Internally iso-8859-1 is used as character set. Other characters
  10  ** are encoded using Unicode entities according to HTML 4.0.
  11  **/
  12
  13    $i18n_php = true;
  14
  15    // Decodes a string to the internal encoding from the given charset
  16    function charset_decode ($charset, $string) {
  17       // All HTML special characters are 7 bit and can be replaced first
  18       $string = htmlspecialchars ($string);
  19
  20       $charset = strtolower($charset);
  21
  22       if (ereg("iso-8859-(.*)", $charset, $res)) {
  23          if ($res[1] == "1")
  24             return charset_decode_iso_8859_1 ($string);
  25          else if ($res[1] == "15")
  26             return charset_decode_iso_8859_15 ($string);
  27          else
  28             return charset_decode_iso_8859_default ($string);
  29       } else if ($charset == "ns_4551-1") {
  30          return charset_decode_ns_4551_1 ($string);
  31       } else
  32          return $string;
  33    }
  34
  35    // iso-8859-1 is the same as Latin 1 and is normally used
  36    // in western europe.
  37    function charset_decode_iso_8859_1 ($string) {
  38       // This is only debug code as long as the internal
  39       // character set is iso-8859-1
  40
  41       // Latin small letter o with stroke
  42       while (ereg("\370", $string))
  43          $string = ereg_replace ("\370", "&#248;", $string);
  44
  45       return ($string);
  46    }
  47
  48    // iso-8859-15 is Latin 15 and has very much the same use as Latin 1
  49    // but has the Euro symbol and some characters needed for French.
  50    function charset_decode_iso_8859_15 ($string) {
  51       // Euro sign
  52       while (ereg("\244", $replace))
  53          $string = ereg_replace ("\244", "&#8364;", $string);
  54       // Latin capital letter S with caron
  55       while (ereg("\246", $string))
  56          $string = ereg_replace ("\244", "&#352;", $string);
  57       // Latin small letter s with caron
  58       while (ereg("\250", $string))
  59          $string = ereg_replace ("\250", "&#353;", $string);
  60       // Latin capital letter Z with caron
  61       while (ereg("\264", $string))
  62          $string = ereg_replace ("\264", "&#381;", $string);
  63       // Latin small letter z with caron
  64       while (ereg("\270", $string))
  65          $string = ereg_replace ("\270", "&#382;", $string);
  66       // Latin capital ligature OE
  67       while (ereg("\274", $string))
  68          $string = ereg_replace ("\274", "&#338;", $string);
  69       // Latin small ligature oe
  70       while (ereg("\275", $string))
  71          $string = ereg_replace ("\275", "&#339;", $string);
  72       // Latin capital letter Y with diaeresis
  73       while (ereg("\276", $string))
  74          $string = ereg_replace ("\276", "&#376;", $string);
  75
  76       return ($string);
  77    }
  78
  79    // Remove al 8 bit characters from all other ISO-8859 character sets
  80    function charset_decode_iso_8859_default ($string) {
  81       return (strtr($string, "\240\241\242\243\244\245\246\247".
  82                     "\250\251\252\253\254\255\256\257".
  83                     "\260\261\262\263\264\265\266\267".
  84                     "\270\271\272\273\274\275\276\277".
  85                     "\300\301\302\303\304\305\306\307".
  86                     "\310\311\312\313\314\315\316\317".
  87                     "\320\321\322\323\324\325\326\327".
  88                     "\330\331\332\333\334\335\336\337".
  89                     "\340\341\342\343\344\345\346\347".
  90                     "\350\351\352\353\354\355\356\357".
  91                     "\360\361\362\363\364\365\366\367".
  92                     "\370\371\372\373\374\375\376\377",
  93                     "????????????????????????????????????????".
  94                     "????????????????????????????????????????".
  95                     "????????????????????????????????????????".
  96                     "????????"));
  97
  98    }
  99
 100    // This is the same as ISO-646-NO and is used by some
 101    // Microsoft programs when sending Norwegian characters
 102    function charset_decode_ns_4551_1 ($string) {
 103       // These characters are:
 104       // Latin capital letter AE
 105       // Latin capital letter O with stroke
 106       // Latin capital letter A with ring above
 107       // and the same as small letters
 108       return strtr ($string, "[\\]{|}", "ÆØÅæøå");
 109    }
 110
 111 ?>